geointel/artifacts/evidence/accuracy/P4/runs/p4-2.0.1-9677d0ef37db82bcf39b/metric-report.json

{
  "case_count": 9,
  "claim_boundary": "Synthetic deterministic reference cases validate evaluator behavior, not production model accuracy.",
  "evaluated_task_families": [
    "change_detection",
    "footprint_segmentation",
    "geospatial_data_validation",
    "object_detection",
    "raster_classification",
    "terrain_interpretation",
    "vector_comparison"
  ],
  "evaluator_version": "2.1.0",
  "failure_taxonomy": {
    "claim_boundary": "diagnostic classification; not a release decision",
    "contexts": {
      "high_confidence": "false positive meets the highest frozen diagnostic confidence threshold",
      "out_of_distribution": "case metadata explicitly marks OOD context",
      "tile_edge": "case metadata explicitly marks tile-edge context"
    },
    "primary_codes": {
      "D-VALIDATION-FN": "reference anomaly or severity not exactly reported",
      "D-VALIDATION-FP": "anomaly or severity reported without an exact reference match",
      "M-AREA-BIAS": "matched footprint has a measurable relative area bias",
      "M-BOUNDARY": "matched footprint has a measurable boundary deviation",
      "M-CLASS": "raster class differs from the valid reference class",
      "M-FN-MISSED": "unmatched reference object or event",
      "M-FP-CONFUSER": "unmatched object or event prediction",
      "M-MISCALIBRATED": "confidence diagnostic deviates from observed correctness",
      "M-OOD": "error observed in an explicitly declared out-of-distribution case",
      "P-PARTIAL": "required prediction value is unavailable"
    }
  },
  "full_results_canonical_json_sha256": "55cdba9cce4dcdc88cee0829cd510d7e624a3472feb9d33f7e681621a792eba3",
  "hash_specification": {
    "algorithm": "SHA-256",
    "canonical_json": "UTF-8 JSON produced with sort_keys=true, separators=(',', ':'), ensure_ascii=false and allow_nan=false",
    "portfolio_canonical_json_sha256_input": "parsed complete portfolio",
    "portfolio_file_sha256_input": "exact source-file bytes",
    "results_canonical_json_sha256_input": "complete sample-id-ordered results array"
  },
  "metric_results_canonical_json_sha256": "cdaf2460bf51231d8f547402ab9c464a72fa9226aa69380e313609f9fb2e3b2e",
  "portfolio_canonical_json_sha256": "1ac1bd77db125db48f9168406d51bfb723522f54a6c81dbe868d600c72f27b17",
  "portfolio_file_sha256": "c49a9c8d74cae2062552b012d9c275f65156c81d75b3e453860f5b859c1c5473",
  "portfolio_id": "geointel-p4-reference-harness-v2",
  "portfolio_kind": "synthetic_contract",
  "portfolio_metrics": {
    "change_detection": {
      "case_support": 1,
      "macro": {
        "f1": 0.5,
        "f1_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "precision": 0.5,
        "precision_case_support": 1,
        "recall": 0.5,
        "recall_case_support": 1,
        "status": "computed"
      },
      "micro": {
        "f1": 0.5,
        "false_discovery_rate": 0.5,
        "false_negative": 1,
        "false_positive": 1,
        "miss_rate": 0.5,
        "precision": 0.5,
        "precision_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "prediction_count": 2,
        "recall": 0.5,
        "recall_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "reference_count": 2,
        "true_positive": 1
      },
      "minimum_case_support": 5,
      "observation_support": {
        "predictions": 2,
        "references": 2
      },
      "primary_metric": {
        "direction": "higher_is_better",
        "name": "micro.f1",
        "value": 0.5
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    },
    "footprint_segmentation": {
      "case_support": 1,
      "macro": {
        "f1": 1.0,
        "f1_case_support": 1,
        "mean_boundary_f1": 1.0,
        "mean_boundary_f1_case_support": 1,
        "mean_dice": 0.8881578947368421,
        "mean_dice_case_support": 1,
        "mean_iou": 0.7988165680473372,
        "mean_iou_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "precision": 1.0,
        "precision_case_support": 1,
        "recall": 1.0,
        "recall_case_support": 1,
        "status": "computed"
      },
      "micro": {
        "f1": 1.0,
        "false_discovery_rate": 0.0,
        "false_negative": 0,
        "false_positive": 0,
        "miss_rate": 0.0,
        "precision": 1.0,
        "precision_ci95_wilson": {
          "lower": 0.20654931437723745,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 1,
          "upper": 1.0
        },
        "prediction_count": 1,
        "recall": 1.0,
        "recall_ci95_wilson": {
          "lower": 0.20654931437723745,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 1,
          "upper": 1.0
        },
        "reference_count": 1,
        "true_positive": 1
      },
      "minimum_case_support": 5,
      "observation_support": {
        "predictions": 1,
        "references": 1
      },
      "primary_metric": {
        "direction": "higher_is_better",
        "name": "micro.f1",
        "value": 1.0
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    },
    "geospatial_data_validation": {
      "case_support": 1,
      "macro": {
        "blocker_or_critical_miss_count": 1.0,
        "blocker_or_critical_miss_count_case_support": 1,
        "f1": 0.6666666666666666,
        "f1_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "precision": 0.6666666666666666,
        "precision_case_support": 1,
        "recall": 0.6666666666666666,
        "recall_case_support": 1,
        "status": "computed"
      },
      "micro": {
        "blocker_or_critical_miss_count": 1,
        "f1": 0.6666666666666666,
        "false_discovery_rate": 0.3333333333333333,
        "false_negative": 1,
        "false_positive": 1,
        "miss_rate": 0.3333333333333333,
        "precision": 0.6666666666666666,
        "precision_ci95_wilson": {
          "lower": 0.2076596008020477,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 3,
          "upper": 0.9385080552796037
        },
        "prediction_count": 3,
        "recall": 0.6666666666666666,
        "recall_ci95_wilson": {
          "lower": 0.2076596008020477,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 3,
          "upper": 0.9385080552796037
        },
        "reference_count": 3,
        "true_positive": 2
      },
      "minimum_case_support": 5,
      "observation_support": {
        "predictions": 3,
        "references": 3
      },
      "primary_metric": {
        "direction": "higher_is_better",
        "name": "micro.f1",
        "value": 0.6666666666666666
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    },
    "object_detection": {
      "case_support": 3,
      "macro": {
        "f1": 0.8,
        "f1_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "precision": 0.3333333333333333,
        "precision_case_support": 2,
        "recall": 1.0,
        "recall_case_support": 1,
        "status": "computed"
      },
      "micro": {
        "ap50": 0.6666666666666666,
        "ap50_95": 0.5838283828382838,
        "calibration": {
          "binning": "five fixed equal-width bins",
          "bins": [
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.0,
              "mean_confidence": null,
              "upper": 0.2
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.2,
              "mean_confidence": null,
              "upper": 0.4
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.4,
              "mean_confidence": null,
              "upper": 0.6
            },
            {
              "accuracy": 0.5,
              "count": 2,
              "lower": 0.6,
              "mean_confidence": 0.6499999999999999,
              "upper": 0.8
            },
            {
              "accuracy": 0.5,
              "count": 2,
              "lower": 0.8,
              "mean_confidence": 0.91,
              "upper": 1.0
            }
          ],
          "brier": 0.3266,
          "ece": 0.27999999999999997,
          "status": "computed"
        },
        "f1": 0.6666666666666666,
        "false_discovery_rate": 0.5,
        "false_negative": 0,
        "false_positive": 2,
        "map50": 0.6666666666666666,
        "map50_95": 0.5838283828382838,
        "miss_rate": 0.0,
        "per_class_ap": {
          "building": {
            "ap50": 0.6666666666666666,
            "ap50_95": 0.5838283828382838,
            "prediction_count": 5,
            "reference_count": 2
          }
        },
        "precision": 0.5,
        "precision_ci95_wilson": {
          "lower": 0.15003898915214947,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 4,
          "upper": 0.8499610108478506
        },
        "prediction_count": 4,
        "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
        "recall": 1.0,
        "recall_ci95_wilson": {
          "lower": 0.34238022750665303,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 1.0
        },
        "reference_count": 2,
        "true_positive": 2
      },
      "minimum_case_support": 5,
      "observation_support": {
        "calibration_predictions": 4,
        "predictions": 4,
        "ranking_predictions": 5,
        "references": 2
      },
      "primary_metric": {
        "direction": "higher_is_better",
        "name": "micro.f1",
        "value": 0.6666666666666666
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    },
    "raster_classification": {
      "case_support": 1,
      "macro": {
        "accuracy": 0.8333333333333334,
        "accuracy_case_support": 1,
        "macro_f1": 0.8222222222222223,
        "macro_f1_case_support": 1,
        "mean_iou": 0.7222222222222222,
        "mean_iou_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "status": "computed"
      },
      "micro": {
        "accuracy": 0.8333333333333334,
        "accuracy_ci95_wilson": {
          "lower": 0.5519691377470266,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 12,
          "upper": 0.9530348578161463
        },
        "correct_pixel_count": 10,
        "evaluated_reference_pixel_count": 12,
        "macro_f1_across_classes": 0.8222222222222223,
        "mean_iou": 0.7222222222222222,
        "per_class": {
          "0": {
            "f1": 1.0,
            "false_discovery_rate": 0.0,
            "false_negative": 0,
            "false_positive": 0,
            "iou": 1.0,
            "miss_rate": 0.0,
            "precision": 1.0,
            "precision_ci95_wilson": {
              "lower": 0.5101091635454027,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 4,
              "upper": 1.0
            },
            "prediction_count": 4,
            "recall": 1.0,
            "recall_ci95_wilson": {
              "lower": 0.5101091635454027,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 4,
              "upper": 1.0
            },
            "reference_count": 4,
            "true_positive": 4
          },
          "1": {
            "f1": 0.8000000000000002,
            "false_discovery_rate": 0.2,
            "false_negative": 1,
            "false_positive": 1,
            "iou": 0.6666666666666666,
            "miss_rate": 0.2,
            "precision": 0.8,
            "precision_ci95_wilson": {
              "lower": 0.37553462976252533,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 5,
              "upper": 0.9637758913675698
            },
            "prediction_count": 5,
            "recall": 0.8,
            "recall_ci95_wilson": {
              "lower": 0.37553462976252533,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 5,
              "upper": 0.9637758913675698
            },
            "reference_count": 5,
            "true_positive": 4
          },
          "2": {
            "f1": 0.6666666666666666,
            "false_discovery_rate": 0.3333333333333333,
            "false_negative": 1,
            "false_positive": 1,
            "iou": 0.5,
            "miss_rate": 0.3333333333333333,
            "precision": 0.6666666666666666,
            "precision_ci95_wilson": {
              "lower": 0.2076596008020477,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 3,
              "upper": 0.9385080552796037
            },
            "prediction_count": 3,
            "recall": 0.6666666666666666,
            "recall_ci95_wilson": {
              "lower": 0.2076596008020477,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 3,
              "upper": 0.9385080552796037
            },
            "reference_count": 3,
            "true_positive": 2
          }
        },
        "prediction_coverage": 1.0,
        "prediction_coverage_ci95_wilson": {
          "lower": 0.7575059933447592,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 12,
          "upper": 1.0
        },
        "prediction_present_pixel_count": 12
      },
      "minimum_case_support": 5,
      "observation_support": {
        "evaluated_pixels": 12
      },
      "primary_metric": {
        "direction": "higher_is_better",
        "name": "micro.mean_iou",
        "value": 0.7222222222222222
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    },
    "terrain_interpretation": {
      "case_support": 1,
      "macro": {
        "bias": 0.02499999999999991,
        "bias_case_support": 1,
        "coverage": 0.8,
        "coverage_case_support": 1,
        "mae": 0.2749999999999999,
        "mae_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "rmse": 0.2872281323269015,
        "rmse_case_support": 1,
        "status": "computed"
      },
      "micro": {
        "bias": 0.02499999999999991,
        "coverage": 0.8,
        "coverage_ci95_wilson": {
          "lower": 0.37553462976252533,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 5,
          "upper": 0.9637758913675698
        },
        "evaluated_count": 4,
        "mae": 0.2749999999999999,
        "reference_count": 5,
        "rmse": 0.2872281323269015,
        "status": "computed",
        "units": "m_TAW"
      },
      "minimum_case_support": 5,
      "observation_support": {
        "evaluated": 4,
        "references": 5
      },
      "primary_metric": {
        "direction": "lower_is_better",
        "name": "micro.rmse",
        "value": 0.2872281323269015
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    },
    "vector_comparison": {
      "case_support": 1,
      "macro": {
        "f1": 0.5,
        "f1_case_support": 1,
        "mean_iou": 1.0,
        "mean_iou_case_support": 1,
        "method": "unweighted mean across cases with defined values",
        "precision": 0.5,
        "precision_case_support": 1,
        "recall": 0.5,
        "recall_case_support": 1,
        "status": "computed"
      },
      "micro": {
        "f1": 0.5,
        "false_discovery_rate": 0.5,
        "false_negative": 1,
        "false_positive": 1,
        "miss_rate": 0.5,
        "precision": 0.5,
        "precision_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "prediction_count": 2,
        "recall": 0.5,
        "recall_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "reference_count": 2,
        "true_positive": 1
      },
      "minimum_case_support": 5,
      "observation_support": {
        "predictions": 2,
        "references": 2
      },
      "primary_metric": {
        "direction": "higher_is_better",
        "name": "micro.f1",
        "value": 0.5
      },
      "release_gate_reason": "insufficient_support",
      "release_gate_status": "not_evaluable",
      "status": "insufficient_support"
    }
  },
  "protected_policy": {
    "diagnostic_curves_select_operating_point": false,
    "operating_point_selection_allowed": false,
    "test_feedback_allowed": false,
    "threshold_selection_source": "pre_registered_configuration_only"
  },
  "results": [
    {
      "metadata": {
        "context": "coastal_hard_negative",
        "date": "2025-07-01",
        "difficulty": "hard",
        "label_review_state": "fixture_reviewed",
        "municipality": "Oostende",
        "object_size": "empty",
        "occlusion": "none",
        "ood": false,
        "region": "flanders",
        "resolution_m": 0.25,
        "season": "summer",
        "sensor": "synthetic-rgb",
        "source": "fixture-orthophoto",
        "tile_edge": true,
        "urbanity": "urban",
        "vegetation": "moderate"
      },
      "metrics": {
        "ap50": null,
        "ap50_95": null,
        "calibration": {
          "binning": "five fixed equal-width bins",
          "bins": [
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.0,
              "mean_confidence": null,
              "upper": 0.2
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.2,
              "mean_confidence": null,
              "upper": 0.4
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.4,
              "mean_confidence": null,
              "upper": 0.6
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.6,
              "mean_confidence": null,
              "upper": 0.8
            },
            {
              "accuracy": 0,
              "count": 1,
              "lower": 0.8,
              "mean_confidence": 0.92,
              "upper": 1.0
            }
          ],
          "brier": 0.8464,
          "ece": 0.92,
          "status": "computed"
        },
        "coverage_risk": [
          {
            "false_negative_count": 0,
            "false_positive_count": 1,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 1,
            "retained_prediction_coverage": 1.0,
            "risk": 1.0,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.0,
            "total_prediction_count": 1
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 1,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 1,
            "retained_prediction_coverage": 1.0,
            "risk": 1.0,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.5,
            "total_prediction_count": 1
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 1,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 1,
            "retained_prediction_coverage": 1.0,
            "risk": 1.0,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.7,
            "total_prediction_count": 1
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 1,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 1,
            "retained_prediction_coverage": 1.0,
            "risk": 1.0,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.9,
            "total_prediction_count": 1
          }
        ],
        "f1": null,
        "false_discovery_rate": 1.0,
        "false_negative": 0,
        "false_positive": 1,
        "map50": null,
        "map50_95": null,
        "match_iou": 0.5,
        "matched_iou": {
          "count": 0,
          "max": null,
          "mean": null,
          "median": null,
          "min": null
        },
        "miss_rate": null,
        "operating_confidence": 0.5,
        "per_class": {
          "building": {
            "ap50": null,
            "ap50_95": null,
            "f1": null,
            "false_discovery_rate": 1.0,
            "false_negative": 0,
            "false_positive": 1,
            "miss_rate": null,
            "precision": 0.0,
            "precision_ci95_wilson": {
              "lower": 0.0,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 1,
              "upper": 0.7934506856227626
            },
            "prediction_count": 1,
            "recall": null,
            "recall_ci95_wilson": {
              "lower": null,
              "status": "undefined",
              "support": 0,
              "upper": null
            },
            "reference_count": 0,
            "true_positive": 0
          }
        },
        "precision": 0.0,
        "precision_ci95_wilson": {
          "lower": 0.0,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 1,
          "upper": 0.7934506856227626
        },
        "prediction_count": 1,
        "recall": null,
        "recall_ci95_wilson": {
          "lower": null,
          "status": "undefined",
          "support": 0,
          "upper": null
        },
        "reference_count": 0,
        "true_positive": 0
      },
      "sample_id": "background-test-hard-negative",
      "task": "object_detection"
    },
    {
      "metadata": {
        "context": "pure_background",
        "date": "2025-02-01",
        "difficulty": "hard",
        "label_review_state": "fixture_reviewed",
        "municipality": "Brussel",
        "object_size": "empty",
        "occlusion": "none",
        "ood": false,
        "region": "brussels",
        "resolution_m": 0.25,
        "season": "winter",
        "sensor": "synthetic-rgb",
        "source": "fixture-orthophoto",
        "tile_edge": false,
        "urbanity": "urban",
        "vegetation": "low"
      },
      "metrics": {
        "ap50": null,
        "ap50_95": null,
        "calibration": {
          "bins": [],
          "brier": null,
          "ece": null,
          "status": "undefined"
        },
        "coverage_risk": [
          {
            "false_negative_count": 0,
            "false_positive_count": 0,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 0,
            "retained_prediction_coverage": null,
            "risk": null,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.0,
            "total_prediction_count": 0
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 0,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 0,
            "retained_prediction_coverage": null,
            "risk": null,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.5,
            "total_prediction_count": 0
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 0,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 0,
            "retained_prediction_coverage": null,
            "risk": null,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.7,
            "total_prediction_count": 0
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 0,
            "matched_reference_count": 0,
            "reference_count": 0,
            "reference_coverage": null,
            "retained_prediction_count": 0,
            "retained_prediction_coverage": null,
            "risk": null,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.9,
            "total_prediction_count": 0
          }
        ],
        "f1": null,
        "false_discovery_rate": null,
        "false_negative": 0,
        "false_positive": 0,
        "map50": null,
        "map50_95": null,
        "match_iou": 0.5,
        "matched_iou": {
          "count": 0,
          "max": null,
          "mean": null,
          "median": null,
          "min": null
        },
        "miss_rate": null,
        "operating_confidence": 0.5,
        "per_class": {
          "building": {
            "ap50": null,
            "ap50_95": null,
            "f1": null,
            "false_discovery_rate": null,
            "false_negative": 0,
            "false_positive": 0,
            "miss_rate": null,
            "precision": null,
            "precision_ci95_wilson": {
              "lower": null,
              "status": "undefined",
              "support": 0,
              "upper": null
            },
            "prediction_count": 0,
            "recall": null,
            "recall_ci95_wilson": {
              "lower": null,
              "status": "undefined",
              "support": 0,
              "upper": null
            },
            "reference_count": 0,
            "true_positive": 0
          }
        },
        "precision": null,
        "precision_ci95_wilson": {
          "lower": null,
          "status": "undefined",
          "support": 0,
          "upper": null
        },
        "prediction_count": 0,
        "recall": null,
        "recall_ci95_wilson": {
          "lower": null,
          "status": "undefined",
          "support": 0,
          "upper": null
        },
        "reference_count": 0,
        "true_positive": 0
      },
      "sample_id": "background-test-pure-empty",
      "task": "object_detection"
    },
    {
      "metadata": {
        "context": "industrial_change",
        "date": "2024-05-01/2025-05-01",
        "difficulty": "hard",
        "label_review_state": "fixture_reviewed",
        "municipality": "Liège",
        "object_size": "mixed",
        "occlusion": "none",
        "ood": false,
        "region": "wallonia",
        "resolution_m": 0.5,
        "season": "multi-date",
        "sensor": "vector",
        "source": "fixture-picc-temporal",
        "tile_edge": false,
        "urbanity": "industrial",
        "vegetation": "low"
      },
      "metrics": {
        "event_metrics": {
          "added": {
            "f1": 1.0,
            "false_discovery_rate": 0.0,
            "false_negative": 0,
            "false_positive": 0,
            "miss_rate": 0.0,
            "precision": 1.0,
            "precision_ci95_wilson": {
              "lower": 0.20654931437723745,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 1,
              "upper": 1.0
            },
            "prediction_count": 1,
            "recall": 1.0,
            "recall_ci95_wilson": {
              "lower": 0.20654931437723745,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 1,
              "upper": 1.0
            },
            "reference_count": 1,
            "true_positive": 1
          },
          "removed": {
            "f1": null,
            "false_discovery_rate": 1.0,
            "false_negative": 1,
            "false_positive": 1,
            "miss_rate": 1.0,
            "precision": 0.0,
            "precision_ci95_wilson": {
              "lower": 0.0,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 1,
              "upper": 0.7934506856227626
            },
            "prediction_count": 1,
            "recall": 0.0,
            "recall_ci95_wilson": {
              "lower": 0.0,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 1,
              "upper": 0.7934506856227626
            },
            "reference_count": 1,
            "true_positive": 0
          }
        },
        "f1": 0.5,
        "false_discovery_rate": 0.5,
        "false_negative": 1,
        "false_positive": 1,
        "miss_rate": 0.5,
        "precision": 0.5,
        "precision_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "prediction_count": 2,
        "recall": 0.5,
        "recall_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "reference_count": 2,
        "true_positive": 1
      },
      "sample_id": "change-test-wallonia-industrial",
      "task": "change_detection"
    },
    {
      "metadata": {
        "context": "dense_urban",
        "date": "2025-06-15",
        "difficulty": "hard",
        "label_review_state": "fixture_reviewed",
        "municipality": "Antwerpen",
        "object_size": "mixed",
        "occlusion": "partial",
        "ood": false,
        "region": "flanders",
        "resolution_m": 0.25,
        "season": "summer",
        "sensor": "synthetic-rgb",
        "source": "fixture-orthophoto",
        "tile_edge": false,
        "urbanity": "urban",
        "vegetation": "moderate"
      },
      "metrics": {
        "ap50": 1.0,
        "ap50_95": 0.900990099009901,
        "calibration": {
          "binning": "five fixed equal-width bins",
          "bins": [
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.0,
              "mean_confidence": null,
              "upper": 0.2
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.2,
              "mean_confidence": null,
              "upper": 0.4
            },
            {
              "accuracy": null,
              "count": 0,
              "lower": 0.4,
              "mean_confidence": null,
              "upper": 0.6
            },
            {
              "accuracy": 0.5,
              "count": 2,
              "lower": 0.6,
              "mean_confidence": 0.6499999999999999,
              "upper": 0.8
            },
            {
              "accuracy": 1,
              "count": 1,
              "lower": 0.8,
              "mean_confidence": 0.9,
              "upper": 1.0
            }
          ],
          "brier": 0.15333333333333332,
          "ece": 0.13333333333333325,
          "status": "computed"
        },
        "coverage_risk": [
          {
            "false_negative_count": 0,
            "false_positive_count": 2,
            "matched_reference_count": 2,
            "reference_count": 2,
            "reference_coverage": 1.0,
            "retained_prediction_count": 4,
            "retained_prediction_coverage": 1.0,
            "risk": 0.5,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.0,
            "total_prediction_count": 4
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 1,
            "matched_reference_count": 2,
            "reference_count": 2,
            "reference_coverage": 1.0,
            "retained_prediction_count": 3,
            "retained_prediction_coverage": 0.75,
            "risk": 0.3333333333333333,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.5,
            "total_prediction_count": 4
          },
          {
            "false_negative_count": 0,
            "false_positive_count": 0,
            "matched_reference_count": 2,
            "reference_count": 2,
            "reference_coverage": 1.0,
            "retained_prediction_count": 2,
            "retained_prediction_coverage": 0.5,
            "risk": 0.0,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.7,
            "total_prediction_count": 4
          },
          {
            "false_negative_count": 1,
            "false_positive_count": 0,
            "matched_reference_count": 1,
            "reference_count": 2,
            "reference_coverage": 0.5,
            "retained_prediction_count": 1,
            "retained_prediction_coverage": 0.25,
            "risk": 0.5,
            "risk_definition": "(false_positive + false_negative) / (true_positive + false_positive + false_negative)",
            "threshold": 0.9,
            "total_prediction_count": 4
          }
        ],
        "f1": 0.8,
        "false_discovery_rate": 0.3333333333333333,
        "false_negative": 0,
        "false_positive": 1,
        "map50": 1.0,
        "map50_95": 0.900990099009901,
        "match_iou": 0.5,
        "matched_iou": {
          "count": 2,
          "max": 1.0,
          "mean": 0.9355555555555555,
          "median": 0.9355555555555555,
          "min": 0.8711111111111111
        },
        "miss_rate": 0.0,
        "operating_confidence": 0.5,
        "per_class": {
          "building": {
            "ap50": 1.0,
            "ap50_95": 0.900990099009901,
            "f1": 0.8,
            "false_discovery_rate": 0.3333333333333333,
            "false_negative": 0,
            "false_positive": 1,
            "miss_rate": 0.0,
            "precision": 0.6666666666666666,
            "precision_ci95_wilson": {
              "lower": 0.2076596008020477,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 3,
              "upper": 0.9385080552796037
            },
            "prediction_count": 3,
            "recall": 1.0,
            "recall_ci95_wilson": {
              "lower": 0.34238022750665303,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 2,
              "upper": 1.0
            },
            "reference_count": 2,
            "true_positive": 2
          }
        },
        "precision": 0.6666666666666666,
        "precision_ci95_wilson": {
          "lower": 0.2076596008020477,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 3,
          "upper": 0.9385080552796037
        },
        "prediction_count": 3,
        "recall": 1.0,
        "recall_ci95_wilson": {
          "lower": 0.34238022750665303,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 1.0
        },
        "reference_count": 2,
        "true_positive": 2
      },
      "sample_id": "det-test-flanders-urban",
      "task": "object_detection"
    },
    {
      "metadata": {
        "context": "dense_urban_raster",
        "date": "2025-10-05",
        "difficulty": "hard",
        "label_review_state": "fixture_reviewed",
        "municipality": "Brussel",
        "object_size": "not_applicable",
        "occlusion": "not_applicable",
        "ood": false,
        "region": "brussels",
        "resolution_m": 1,
        "season": "autumn",
        "sensor": "synthetic-multispectral",
        "source": "fixture-thematic-raster",
        "tile_edge": true,
        "urbanity": "urban",
        "vegetation": "moderate"
      },
      "metrics": {
        "accuracy": 0.8333333333333334,
        "accuracy_ci95_wilson": {
          "lower": 0.5519691377470266,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 12,
          "upper": 0.9530348578161463
        },
        "alignment_validated": true,
        "confusion_matrix": {
          "0": {
            "0": 4,
            "1": 0,
            "2": 0,
            "__prediction_nodata__": 0
          },
          "1": {
            "0": 0,
            "1": 4,
            "2": 1,
            "__prediction_nodata__": 0
          },
          "2": {
            "0": 0,
            "1": 1,
            "2": 2,
            "__prediction_nodata__": 0
          }
        },
        "correct_pixel_count": 10,
        "evaluated_reference_pixel_count": 12,
        "ignored_reference_pixel_count": 0,
        "macro_f1": 0.8222222222222223,
        "mean_iou": 0.7222222222222222,
        "per_class": {
          "0": {
            "f1": 1.0,
            "false_discovery_rate": 0.0,
            "false_negative": 0,
            "false_positive": 0,
            "iou": 1.0,
            "miss_rate": 0.0,
            "precision": 1.0,
            "precision_ci95_wilson": {
              "lower": 0.5101091635454027,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 4,
              "upper": 1.0
            },
            "prediction_count": 4,
            "recall": 1.0,
            "recall_ci95_wilson": {
              "lower": 0.5101091635454027,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 4,
              "upper": 1.0
            },
            "reference_count": 4,
            "true_positive": 4
          },
          "1": {
            "f1": 0.8000000000000002,
            "false_discovery_rate": 0.2,
            "false_negative": 1,
            "false_positive": 1,
            "iou": 0.6666666666666666,
            "miss_rate": 0.2,
            "precision": 0.8,
            "precision_ci95_wilson": {
              "lower": 0.37553462976252533,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 5,
              "upper": 0.9637758913675698
            },
            "prediction_count": 5,
            "recall": 0.8,
            "recall_ci95_wilson": {
              "lower": 0.37553462976252533,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 5,
              "upper": 0.9637758913675698
            },
            "reference_count": 5,
            "true_positive": 4
          },
          "2": {
            "f1": 0.6666666666666666,
            "false_discovery_rate": 0.3333333333333333,
            "false_negative": 1,
            "false_positive": 1,
            "iou": 0.5,
            "miss_rate": 0.3333333333333333,
            "precision": 0.6666666666666666,
            "precision_ci95_wilson": {
              "lower": 0.2076596008020477,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 3,
              "upper": 0.9385080552796037
            },
            "prediction_count": 3,
            "recall": 0.6666666666666666,
            "recall_ci95_wilson": {
              "lower": 0.2076596008020477,
              "method": "Wilson score interval",
              "status": "computed",
              "support": 3,
              "upper": 0.9385080552796037
            },
            "reference_count": 3,
            "true_positive": 2
          }
        },
        "pixel_count": 12,
        "prediction_coverage": 1.0,
        "prediction_coverage_ci95_wilson": {
          "lower": 0.7575059933447592,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 12,
          "upper": 1.0
        },
        "prediction_outside_reference_count": 0,
        "prediction_present_pixel_count": 12
      },
      "sample_id": "raster-test-brussels-urban",
      "task": "raster_classification"
    },
    {
      "metadata": {
        "context": "rural_buildings",
        "date": "2025-04-20",
        "difficulty": "normal",
        "label_review_state": "fixture_reviewed",
        "municipality": "Namur",
        "object_size": "large",
        "occlusion": "none",
        "ood": false,
        "region": "wallonia",
        "resolution_m": 0.5,
        "season": "spring",
        "sensor": "synthetic-rgb",
        "source": "fixture-picc",
        "tile_edge": false,
        "urbanity": "rural",
        "vegetation": "low"
      },
      "metrics": {
        "centroid_distance_m": {
          "count": 1,
          "max": 0.5,
          "mean": 0.5,
          "median": 0.5,
          "min": 0.5
        },
        "f1": 1.0,
        "false_discovery_rate": 0.0,
        "false_negative": 0,
        "false_positive": 0,
        "mean_boundary_f1": 1.0,
        "mean_dice": 0.8881578947368421,
        "mean_iou": 0.7988165680473372,
        "miss_rate": 0.0,
        "precision": 1.0,
        "precision_ci95_wilson": {
          "lower": 0.20654931437723745,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 1,
          "upper": 1.0
        },
        "prediction_count": 1,
        "recall": 1.0,
        "recall_ci95_wilson": {
          "lower": 0.20654931437723745,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 1,
          "upper": 1.0
        },
        "reference_count": 1,
        "relative_area_error": {
          "count": 1,
          "max": -0.1,
          "mean": -0.1,
          "median": -0.1,
          "min": -0.1
        },
        "spatial_context_validated": true,
        "topologically_invalid_predictions": 0,
        "topologically_valid_predictions": 1,
        "true_positive": 1
      },
      "sample_id": "seg-test-wallonia-rural",
      "task": "footprint_segmentation"
    },
    {
      "metadata": {
        "context": "terrain_profile",
        "date": "2023-01-01",
        "difficulty": "normal",
        "label_review_state": "fixture_reviewed",
        "municipality": "Hasselt",
        "object_size": "not_applicable",
        "occlusion": "not_applicable",
        "ood": false,
        "region": "flanders",
        "resolution_m": 5,
        "season": "not_applicable",
        "sensor": "elevation-raster",
        "source": "fixture-dhmv",
        "tile_edge": false,
        "urbanity": "rural",
        "vegetation": "moderate"
      },
      "metrics": {
        "absolute_error_sum": 1.0999999999999996,
        "bias": 0.02499999999999991,
        "coverage": 0.8,
        "coverage_ci95_wilson": {
          "lower": 0.37553462976252533,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 5,
          "upper": 0.9637758913675698
        },
        "error_distribution": {
          "count": 4,
          "max": 0.40000000000000036,
          "mean": 0.02499999999999991,
          "median": 0.0,
          "min": -0.3000000000000007
        },
        "error_sum": 0.09999999999999964,
        "evaluated_count": 4,
        "mae": 0.2749999999999999,
        "missing_count": 1,
        "reference_count": 5,
        "rmse": 0.2872281323269015,
        "squared_error_sum": 0.3300000000000001,
        "units": "m_TAW"
      },
      "sample_id": "terrain-test-flanders-rural",
      "task": "terrain_interpretation"
    },
    {
      "metadata": {
        "context": "data_contract",
        "date": "2025-01-01",
        "difficulty": "hard",
        "label_review_state": "fixture_reviewed",
        "municipality": "not_applicable",
        "object_size": "not_applicable",
        "occlusion": "not_applicable",
        "ood": false,
        "region": "national",
        "resolution_m": null,
        "season": "not_applicable",
        "sensor": "mixed",
        "source": "fixture-contracts",
        "tile_edge": false,
        "urbanity": "mixed",
        "vegetation": "mixed"
      },
      "metrics": {
        "blocker_or_critical_miss_count": 1,
        "f1": 0.6666666666666666,
        "false_discovery_rate": 0.3333333333333333,
        "false_negative": 1,
        "false_positive": 1,
        "miss_rate": 0.3333333333333333,
        "misses_by_severity": {
          "blocker": 0,
          "critical": 1,
          "informational": 0,
          "major": 0,
          "minor": 0
        },
        "precision": 0.6666666666666666,
        "precision_ci95_wilson": {
          "lower": 0.2076596008020477,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 3,
          "upper": 0.9385080552796037
        },
        "prediction_count": 3,
        "recall": 0.6666666666666666,
        "recall_ci95_wilson": {
          "lower": 0.2076596008020477,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 3,
          "upper": 0.9385080552796037
        },
        "reference_count": 3,
        "severity_mismatch_count": 0,
        "severity_mismatches": [],
        "true_positive": 2
      },
      "sample_id": "validation-test-national",
      "task": "geospatial_data_validation"
    },
    {
      "metadata": {
        "context": "suburban_buildings",
        "date": "2025-01-15",
        "difficulty": "normal",
        "label_review_state": "fixture_reviewed",
        "municipality": "Mol",
        "object_size": "mixed",
        "occlusion": "not_applicable",
        "ood": false,
        "region": "flanders",
        "resolution_m": 0.1,
        "season": "not_applicable",
        "sensor": "vector",
        "source": "fixture-grb",
        "tile_edge": false,
        "urbanity": "suburban",
        "vegetation": "low"
      },
      "metrics": {
        "f1": 0.5,
        "false_discovery_rate": 0.5,
        "false_negative": 1,
        "false_positive": 1,
        "mean_iou": 1.0,
        "miss_rate": 0.5,
        "precision": 0.5,
        "precision_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "prediction_count": 2,
        "recall": 0.5,
        "recall_ci95_wilson": {
          "lower": 0.09453120573423074,
          "method": "Wilson score interval",
          "status": "computed",
          "support": 2,
          "upper": 0.9054687942657693
        },
        "reference_count": 2,
        "spatial_context_validated": true,
        "topologically_valid": true,
        "true_positive": 1
      },
      "sample_id": "vector-test-flanders-suburban",
      "task": "vector_comparison"
    }
  ],
  "results_canonical_json_sha256": "55cdba9cce4dcdc88cee0829cd510d7e624a3472feb9d33f7e681621a792eba3",
  "schema_version": 2,
  "selection_policy": "No data-dependent threshold, model or gate selection is performed from protected cases.",
  "split_roles": [
    "test",
    "background-test"
  ],
  "subgroups": {
    "dimensions": {
      "context": {
        "strata": {
          "coastal_hard_negative": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.0,
                  "precision_case_support": 1,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "computed"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.6,
                        "mean_confidence": null,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.92,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.8464,
                    "ece": 0.92,
                    "status": "computed"
                  },
                  "f1": null,
                  "false_discovery_rate": 1.0,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 1,
                      "reference_count": 0
                    }
                  },
                  "precision": 0.0,
                  "precision_ci95_wilson": {
                    "lower": 0.0,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 0.7934506856227626
                  },
                  "prediction_count": 1,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 1,
                  "predictions": 1,
                  "ranking_predictions": 1,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "data_contract": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "dense_urban": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 1.0,
                  "ap50_95": 0.900990099009901,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 1,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.9,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.15333333333333332,
                    "ece": 0.13333333333333325,
                    "status": "computed"
                  },
                  "f1": 0.8,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": 1.0,
                  "map50_95": 0.900990099009901,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 1.0,
                      "ap50_95": 0.900990099009901,
                      "prediction_count": 4,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 3,
                  "predictions": 3,
                  "ranking_predictions": 4,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.8
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "dense_urban_raster": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "raster_classification"
            ]
          },
          "industrial_change": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection"
            ]
          },
          "pure_background": {
            "case_support": 1,
            "failure_count": 0,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": null,
                  "precision_case_support": 0,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "not_evaluable"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "bins": [],
                    "brier": null,
                    "ece": null,
                    "status": "undefined"
                  },
                  "f1": null,
                  "false_discovery_rate": null,
                  "false_negative": 0,
                  "false_positive": 0,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 0,
                      "reference_count": 0
                    }
                  },
                  "precision": null,
                  "precision_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "prediction_count": 0,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 0,
                  "predictions": 0,
                  "ranking_predictions": 0,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "rural_buildings": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation"
            ]
          },
          "suburban_buildings": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "vector_comparison"
            ]
          },
          "terrain_profile": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "terrain_interpretation"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "date": {
        "strata": {
          "2023-01-01": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "terrain_interpretation"
            ]
          },
          "2024-05-01/2025-05-01": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection"
            ]
          },
          "2025-01-01": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "2025-01-15": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "vector_comparison"
            ]
          },
          "2025-02-01": {
            "case_support": 1,
            "failure_count": 0,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": null,
                  "precision_case_support": 0,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "not_evaluable"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "bins": [],
                    "brier": null,
                    "ece": null,
                    "status": "undefined"
                  },
                  "f1": null,
                  "false_discovery_rate": null,
                  "false_negative": 0,
                  "false_positive": 0,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 0,
                      "reference_count": 0
                    }
                  },
                  "precision": null,
                  "precision_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "prediction_count": 0,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 0,
                  "predictions": 0,
                  "ranking_predictions": 0,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "2025-04-20": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation"
            ]
          },
          "2025-06-15": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 1.0,
                  "ap50_95": 0.900990099009901,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 1,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.9,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.15333333333333332,
                    "ece": 0.13333333333333325,
                    "status": "computed"
                  },
                  "f1": 0.8,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": 1.0,
                  "map50_95": 0.900990099009901,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 1.0,
                      "ap50_95": 0.900990099009901,
                      "prediction_count": 4,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 3,
                  "predictions": 3,
                  "ranking_predictions": 4,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.8
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "2025-07-01": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.0,
                  "precision_case_support": 1,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "computed"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.6,
                        "mean_confidence": null,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.92,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.8464,
                    "ece": 0.92,
                    "status": "computed"
                  },
                  "f1": null,
                  "false_discovery_rate": 1.0,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 1,
                      "reference_count": 0
                    }
                  },
                  "precision": 0.0,
                  "precision_ci95_wilson": {
                    "lower": 0.0,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 0.7934506856227626
                  },
                  "prediction_count": 1,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 1,
                  "predictions": 1,
                  "ranking_predictions": 1,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "2025-10-05": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "raster_classification"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "difficulty": {
        "strata": {
          "hard": {
            "case_support": 6,
            "failure_count": 10,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "object_detection": {
                "case_support": 3,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "geospatial_data_validation",
              "object_detection",
              "raster_classification"
            ]
          },
          "normal": {
            "case_support": 3,
            "failure_count": 4,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation",
              "terrain_interpretation",
              "vector_comparison"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "municipality": {
        "strata": {
          "Antwerpen": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 1.0,
                  "ap50_95": 0.900990099009901,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 1,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.9,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.15333333333333332,
                    "ece": 0.13333333333333325,
                    "status": "computed"
                  },
                  "f1": 0.8,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": 1.0,
                  "map50_95": 0.900990099009901,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 1.0,
                      "ap50_95": 0.900990099009901,
                      "prediction_count": 4,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 3,
                  "predictions": 3,
                  "ranking_predictions": 4,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.8
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "Brussel": {
            "case_support": 2,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": null,
                  "precision_case_support": 0,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "not_evaluable"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "bins": [],
                    "brier": null,
                    "ece": null,
                    "status": "undefined"
                  },
                  "f1": null,
                  "false_discovery_rate": null,
                  "false_negative": 0,
                  "false_positive": 0,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 0,
                      "reference_count": 0
                    }
                  },
                  "precision": null,
                  "precision_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "prediction_count": 0,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 0,
                  "predictions": 0,
                  "ranking_predictions": 0,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection",
              "raster_classification"
            ]
          },
          "Hasselt": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "terrain_interpretation"
            ]
          },
          "Liège": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection"
            ]
          },
          "Mol": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "vector_comparison"
            ]
          },
          "Namur": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation"
            ]
          },
          "Oostende": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.0,
                  "precision_case_support": 1,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "computed"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.6,
                        "mean_confidence": null,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.92,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.8464,
                    "ece": 0.92,
                    "status": "computed"
                  },
                  "f1": null,
                  "false_discovery_rate": 1.0,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 1,
                      "reference_count": 0
                    }
                  },
                  "precision": 0.0,
                  "precision_ci95_wilson": {
                    "lower": 0.0,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 0.7934506856227626
                  },
                  "prediction_count": 1,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 1,
                  "predictions": 1,
                  "ranking_predictions": 1,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "not_applicable": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "object_size": {
        "strata": {
          "empty": {
            "case_support": 2,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 2,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.0,
                  "precision_case_support": 1,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "computed"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.6,
                        "mean_confidence": null,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.92,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.8464,
                    "ece": 0.92,
                    "status": "computed"
                  },
                  "f1": null,
                  "false_discovery_rate": 1.0,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 1,
                      "reference_count": 0
                    }
                  },
                  "precision": 0.0,
                  "precision_ci95_wilson": {
                    "lower": 0.0,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 0.7934506856227626
                  },
                  "prediction_count": 1,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 1,
                  "predictions": 1,
                  "ranking_predictions": 1,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "large": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation"
            ]
          },
          "mixed": {
            "case_support": 3,
            "failure_count": 6,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 1.0,
                  "ap50_95": 0.900990099009901,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 1,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.9,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.15333333333333332,
                    "ece": 0.13333333333333325,
                    "status": "computed"
                  },
                  "f1": 0.8,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": 1.0,
                  "map50_95": 0.900990099009901,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 1.0,
                      "ap50_95": 0.900990099009901,
                      "prediction_count": 4,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 3,
                  "predictions": 3,
                  "ranking_predictions": 4,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.8
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "object_detection",
              "vector_comparison"
            ]
          },
          "not_applicable": {
            "case_support": 3,
            "failure_count": 5,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation",
              "raster_classification",
              "terrain_interpretation"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "occlusion": {
        "strata": {
          "none": {
            "case_support": 4,
            "failure_count": 5,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "object_detection": {
                "case_support": 2,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.0,
                  "precision_case_support": 1,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "computed"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.6,
                        "mean_confidence": null,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.92,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.8464,
                    "ece": 0.92,
                    "status": "computed"
                  },
                  "f1": null,
                  "false_discovery_rate": 1.0,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 1,
                      "reference_count": 0
                    }
                  },
                  "precision": 0.0,
                  "precision_ci95_wilson": {
                    "lower": 0.0,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 0.7934506856227626
                  },
                  "prediction_count": 1,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 1,
                  "predictions": 1,
                  "ranking_predictions": 1,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "footprint_segmentation",
              "object_detection"
            ]
          },
          "not_applicable": {
            "case_support": 4,
            "failure_count": 7,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation",
              "raster_classification",
              "terrain_interpretation",
              "vector_comparison"
            ]
          },
          "partial": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 1.0,
                  "ap50_95": 0.900990099009901,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 1,
                        "count": 1,
                        "lower": 0.8,
                        "mean_confidence": 0.9,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.15333333333333332,
                    "ece": 0.13333333333333325,
                    "status": "computed"
                  },
                  "f1": 0.8,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 0,
                  "false_positive": 1,
                  "map50": 1.0,
                  "map50_95": 0.900990099009901,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 1.0,
                      "ap50_95": 0.900990099009901,
                      "prediction_count": 4,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 3,
                  "predictions": 3,
                  "ranking_predictions": 4,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.8
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "region": {
        "strata": {
          "brussels": {
            "case_support": 2,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": null,
                  "precision_case_support": 0,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "not_evaluable"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "bins": [],
                    "brier": null,
                    "ece": null,
                    "status": "undefined"
                  },
                  "f1": null,
                  "false_discovery_rate": null,
                  "false_negative": 0,
                  "false_positive": 0,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 0,
                      "reference_count": 0
                    }
                  },
                  "precision": null,
                  "precision_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "prediction_count": 0,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 0,
                  "predictions": 0,
                  "ranking_predictions": 0,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection",
              "raster_classification"
            ]
          },
          "flanders": {
            "case_support": 4,
            "failure_count": 7,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 2,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection",
              "terrain_interpretation",
              "vector_comparison"
            ]
          },
          "national": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "wallonia": {
            "case_support": 2,
            "failure_count": 3,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "footprint_segmentation"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "resolution_m": {
        "strata": {
          "0.1": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "vector_comparison"
            ]
          },
          "0.25": {
            "case_support": 3,
            "failure_count": 4,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 3,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "0.5": {
            "case_support": 2,
            "failure_count": 3,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "footprint_segmentation"
            ]
          },
          "1": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "raster_classification"
            ]
          },
          "5": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "terrain_interpretation"
            ]
          },
          "__not_applicable__": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "season": {
        "strata": {
          "autumn": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "raster_classification"
            ]
          },
          "multi-date": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection"
            ]
          },
          "not_applicable": {
            "case_support": 3,
            "failure_count": 5,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation",
              "terrain_interpretation",
              "vector_comparison"
            ]
          },
          "spring": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation"
            ]
          },
          "summer": {
            "case_support": 2,
            "failure_count": 4,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 2,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "winter": {
            "case_support": 1,
            "failure_count": 0,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": null,
                  "precision_case_support": 0,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "not_evaluable"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "bins": [],
                    "brier": null,
                    "ece": null,
                    "status": "undefined"
                  },
                  "f1": null,
                  "false_discovery_rate": null,
                  "false_negative": 0,
                  "false_positive": 0,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 0,
                      "reference_count": 0
                    }
                  },
                  "precision": null,
                  "precision_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "prediction_count": 0,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 0,
                  "predictions": 0,
                  "ranking_predictions": 0,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "sensor": {
        "strata": {
          "elevation-raster": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "terrain_interpretation"
            ]
          },
          "mixed": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "synthetic-multispectral": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "raster_classification"
            ]
          },
          "synthetic-rgb": {
            "case_support": 4,
            "failure_count": 5,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "object_detection": {
                "case_support": 3,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation",
              "object_detection"
            ]
          },
          "vector": {
            "case_support": 2,
            "failure_count": 4,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "vector_comparison"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "source": {
        "strata": {
          "fixture-contracts": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "fixture-dhmv": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "terrain_interpretation"
            ]
          },
          "fixture-grb": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "vector_comparison"
            ]
          },
          "fixture-orthophoto": {
            "case_support": 3,
            "failure_count": 4,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 3,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection"
            ]
          },
          "fixture-picc": {
            "case_support": 1,
            "failure_count": 1,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation"
            ]
          },
          "fixture-picc-temporal": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection"
            ]
          },
          "fixture-thematic-raster": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "raster_classification"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "urbanity": {
        "strata": {
          "industrial": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection"
            ]
          },
          "mixed": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "rural": {
            "case_support": 2,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "footprint_segmentation",
              "terrain_interpretation"
            ]
          },
          "suburban": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "vector_comparison"
            ]
          },
          "urban": {
            "case_support": 4,
            "failure_count": 6,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 3,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection",
              "raster_classification"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      },
      "vegetation": {
        "strata": {
          "low": {
            "case_support": 4,
            "failure_count": 5,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "change_detection": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "footprint_segmentation": {
                "case_support": 1,
                "macro": {
                  "f1": 1.0,
                  "f1_case_support": 1,
                  "mean_boundary_f1": 1.0,
                  "mean_boundary_f1_case_support": 1,
                  "mean_dice": 0.8881578947368421,
                  "mean_dice_case_support": 1,
                  "mean_iou": 0.7988165680473372,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 1.0,
                  "precision_case_support": 1,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 1.0,
                  "false_discovery_rate": 0.0,
                  "false_negative": 0,
                  "false_positive": 0,
                  "miss_rate": 0.0,
                  "precision": 1.0,
                  "precision_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "prediction_count": 1,
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.20654931437723745,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 1,
                    "upper": 1.0
                  },
                  "reference_count": 1,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 1,
                  "references": 1
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 1.0
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "object_detection": {
                "case_support": 1,
                "macro": {
                  "f1": null,
                  "f1_case_support": 0,
                  "method": "unweighted mean across cases with defined values",
                  "precision": null,
                  "precision_case_support": 0,
                  "recall": null,
                  "recall_case_support": 0,
                  "status": "not_evaluable"
                },
                "micro": {
                  "ap50": null,
                  "ap50_95": null,
                  "calibration": {
                    "bins": [],
                    "brier": null,
                    "ece": null,
                    "status": "undefined"
                  },
                  "f1": null,
                  "false_discovery_rate": null,
                  "false_negative": 0,
                  "false_positive": 0,
                  "map50": null,
                  "map50_95": null,
                  "miss_rate": null,
                  "per_class_ap": {
                    "building": {
                      "ap50": null,
                      "ap50_95": null,
                      "prediction_count": 0,
                      "reference_count": 0
                    }
                  },
                  "precision": null,
                  "precision_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "prediction_count": 0,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": null,
                  "recall_ci95_wilson": {
                    "lower": null,
                    "status": "undefined",
                    "support": 0,
                    "upper": null
                  },
                  "reference_count": 0,
                  "true_positive": 0
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 0,
                  "predictions": 0,
                  "ranking_predictions": 0,
                  "references": 0
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": null
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "vector_comparison": {
                "case_support": 1,
                "macro": {
                  "f1": 0.5,
                  "f1_case_support": 1,
                  "mean_iou": 1.0,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.5,
                  "precision_case_support": 1,
                  "recall": 0.5,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "f1": 0.5,
                  "false_discovery_rate": 0.5,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.5,
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "prediction_count": 2,
                  "recall": 0.5,
                  "recall_ci95_wilson": {
                    "lower": 0.09453120573423074,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 0.9054687942657693
                  },
                  "reference_count": 2,
                  "true_positive": 1
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 2,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.5
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "change_detection",
              "footprint_segmentation",
              "object_detection",
              "vector_comparison"
            ]
          },
          "mixed": {
            "case_support": 1,
            "failure_count": 2,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "geospatial_data_validation": {
                "case_support": 1,
                "macro": {
                  "blocker_or_critical_miss_count": 1.0,
                  "blocker_or_critical_miss_count_case_support": 1,
                  "f1": 0.6666666666666666,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.6666666666666666,
                  "precision_case_support": 1,
                  "recall": 0.6666666666666666,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "blocker_or_critical_miss_count": 1,
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.3333333333333333,
                  "false_negative": 1,
                  "false_positive": 1,
                  "miss_rate": 0.3333333333333333,
                  "precision": 0.6666666666666666,
                  "precision_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "prediction_count": 3,
                  "recall": 0.6666666666666666,
                  "recall_ci95_wilson": {
                    "lower": 0.2076596008020477,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 3,
                    "upper": 0.9385080552796037
                  },
                  "reference_count": 3,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "predictions": 3,
                  "references": 3
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "geospatial_data_validation"
            ]
          },
          "moderate": {
            "case_support": 4,
            "failure_count": 7,
            "release_gate_reason": "insufficient_support",
            "release_gate_status": "not_evaluable",
            "task_metrics": {
              "object_detection": {
                "case_support": 2,
                "macro": {
                  "f1": 0.8,
                  "f1_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "precision": 0.3333333333333333,
                  "precision_case_support": 2,
                  "recall": 1.0,
                  "recall_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "ap50": 0.6666666666666666,
                  "ap50_95": 0.5838283828382838,
                  "calibration": {
                    "binning": "five fixed equal-width bins",
                    "bins": [
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.0,
                        "mean_confidence": null,
                        "upper": 0.2
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.2,
                        "mean_confidence": null,
                        "upper": 0.4
                      },
                      {
                        "accuracy": null,
                        "count": 0,
                        "lower": 0.4,
                        "mean_confidence": null,
                        "upper": 0.6
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.6,
                        "mean_confidence": 0.6499999999999999,
                        "upper": 0.8
                      },
                      {
                        "accuracy": 0.5,
                        "count": 2,
                        "lower": 0.8,
                        "mean_confidence": 0.91,
                        "upper": 1.0
                      }
                    ],
                    "brier": 0.3266,
                    "ece": 0.27999999999999997,
                    "status": "computed"
                  },
                  "f1": 0.6666666666666666,
                  "false_discovery_rate": 0.5,
                  "false_negative": 0,
                  "false_positive": 2,
                  "map50": 0.6666666666666666,
                  "map50_95": 0.5838283828382838,
                  "miss_rate": 0.0,
                  "per_class_ap": {
                    "building": {
                      "ap50": 0.6666666666666666,
                      "ap50_95": 0.5838283828382838,
                      "prediction_count": 5,
                      "reference_count": 2
                    }
                  },
                  "precision": 0.5,
                  "precision_ci95_wilson": {
                    "lower": 0.15003898915214947,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 4,
                    "upper": 0.8499610108478506
                  },
                  "prediction_count": 4,
                  "ranking_scope": "predictions pooled across cases; class- and sample-aware matching; no averaging of per-case AP",
                  "recall": 1.0,
                  "recall_ci95_wilson": {
                    "lower": 0.34238022750665303,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 2,
                    "upper": 1.0
                  },
                  "reference_count": 2,
                  "true_positive": 2
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "calibration_predictions": 4,
                  "predictions": 4,
                  "ranking_predictions": 5,
                  "references": 2
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.f1",
                  "value": 0.6666666666666666
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "raster_classification": {
                "case_support": 1,
                "macro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_case_support": 1,
                  "macro_f1": 0.8222222222222223,
                  "macro_f1_case_support": 1,
                  "mean_iou": 0.7222222222222222,
                  "mean_iou_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "status": "computed"
                },
                "micro": {
                  "accuracy": 0.8333333333333334,
                  "accuracy_ci95_wilson": {
                    "lower": 0.5519691377470266,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 0.9530348578161463
                  },
                  "correct_pixel_count": 10,
                  "evaluated_reference_pixel_count": 12,
                  "macro_f1_across_classes": 0.8222222222222223,
                  "mean_iou": 0.7222222222222222,
                  "per_class": {
                    "0": {
                      "f1": 1.0,
                      "false_discovery_rate": 0.0,
                      "false_negative": 0,
                      "false_positive": 0,
                      "iou": 1.0,
                      "miss_rate": 0.0,
                      "precision": 1.0,
                      "precision_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "prediction_count": 4,
                      "recall": 1.0,
                      "recall_ci95_wilson": {
                        "lower": 0.5101091635454027,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 4,
                        "upper": 1.0
                      },
                      "reference_count": 4,
                      "true_positive": 4
                    },
                    "1": {
                      "f1": 0.8000000000000002,
                      "false_discovery_rate": 0.2,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.6666666666666666,
                      "miss_rate": 0.2,
                      "precision": 0.8,
                      "precision_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "prediction_count": 5,
                      "recall": 0.8,
                      "recall_ci95_wilson": {
                        "lower": 0.37553462976252533,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 5,
                        "upper": 0.9637758913675698
                      },
                      "reference_count": 5,
                      "true_positive": 4
                    },
                    "2": {
                      "f1": 0.6666666666666666,
                      "false_discovery_rate": 0.3333333333333333,
                      "false_negative": 1,
                      "false_positive": 1,
                      "iou": 0.5,
                      "miss_rate": 0.3333333333333333,
                      "precision": 0.6666666666666666,
                      "precision_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "prediction_count": 3,
                      "recall": 0.6666666666666666,
                      "recall_ci95_wilson": {
                        "lower": 0.2076596008020477,
                        "method": "Wilson score interval",
                        "status": "computed",
                        "support": 3,
                        "upper": 0.9385080552796037
                      },
                      "reference_count": 3,
                      "true_positive": 2
                    }
                  },
                  "prediction_coverage": 1.0,
                  "prediction_coverage_ci95_wilson": {
                    "lower": 0.7575059933447592,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 12,
                    "upper": 1.0
                  },
                  "prediction_present_pixel_count": 12
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated_pixels": 12
                },
                "primary_metric": {
                  "direction": "higher_is_better",
                  "name": "micro.mean_iou",
                  "value": 0.7222222222222222
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              },
              "terrain_interpretation": {
                "case_support": 1,
                "macro": {
                  "bias": 0.02499999999999991,
                  "bias_case_support": 1,
                  "coverage": 0.8,
                  "coverage_case_support": 1,
                  "mae": 0.2749999999999999,
                  "mae_case_support": 1,
                  "method": "unweighted mean across cases with defined values",
                  "rmse": 0.2872281323269015,
                  "rmse_case_support": 1,
                  "status": "computed"
                },
                "micro": {
                  "bias": 0.02499999999999991,
                  "coverage": 0.8,
                  "coverage_ci95_wilson": {
                    "lower": 0.37553462976252533,
                    "method": "Wilson score interval",
                    "status": "computed",
                    "support": 5,
                    "upper": 0.9637758913675698
                  },
                  "evaluated_count": 4,
                  "mae": 0.2749999999999999,
                  "reference_count": 5,
                  "rmse": 0.2872281323269015,
                  "status": "computed",
                  "units": "m_TAW"
                },
                "minimum_case_support": 5,
                "observation_support": {
                  "evaluated": 4,
                  "references": 5
                },
                "primary_metric": {
                  "direction": "lower_is_better",
                  "name": "micro.rmse",
                  "value": 0.2872281323269015
                },
                "release_gate_reason": "insufficient_support",
                "release_gate_status": "not_evaluable",
                "status": "insufficient_support"
              }
            },
            "tasks": [
              "object_detection",
              "raster_classification",
              "terrain_interpretation"
            ]
          }
        },
        "worst_stratum_by_task": {
          "change_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "footprint_segmentation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "geospatial_data_validation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "object_detection": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "raster_classification": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "terrain_interpretation": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          },
          "vector_comparison": {
            "reason": "no stratum meets minimum support with a defined primary metric",
            "status": "not_evaluable"
          }
        }
      }
    },
    "minimum_case_support": 5,
    "overall_status": "not_evaluable",
    "schema_version": 2,
    "support_unit": "independent benchmark cases; pixel/object counts are reported separately and do not replace case support"
  },
  "task_count": 7,
  "task_inventory": [
    {
      "capability_id": "model_object_detection",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_contract_case_only",
      "implementation_kind": "model_inference_pipeline",
      "implementation_paths": [
        "backend/app/services/detection_service.py"
      ],
      "suitable_metrics": [
        "precision",
        "recall",
        "F1",
        "AP50",
        "AP50-95",
        "IoU",
        "ECE",
        "Brier",
        "coverage-risk"
      ],
      "task": "object_detection"
    },
    {
      "capability_id": "building_proposal_filtering",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "not_separately_benchmarked",
      "implementation_kind": "supporting_candidate_classifier",
      "implementation_paths": [
        "scripts/train_building_proposal_classifier.py",
        "scripts/evaluate_belgium_building_candidate.py"
      ],
      "suitable_metrics": [
        "candidate precision",
        "candidate recall",
        "F1",
        "calibration"
      ],
      "task": "object_detection"
    },
    {
      "capability_id": "footprint_segmentation",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_contract_case_only",
      "implementation_kind": "model_inference_pipeline",
      "implementation_paths": [
        "backend/app/services/segmentation_service.py",
        "backend/app/services/segmentation_adapter.py"
      ],
      "suitable_metrics": [
        "object precision",
        "object recall",
        "F1",
        "IoU",
        "Dice",
        "boundary F1",
        "centroid distance",
        "area error",
        "topology"
      ],
      "task": "footprint_segmentation"
    },
    {
      "capability_id": "detection_and_vector_qa",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_contract_case_only",
      "implementation_kind": "deterministic_geospatial_comparison",
      "implementation_paths": [
        "backend/app/services/qa_service.py",
        "backend/app/services/detection_qa_service.py",
        "backend/app/services/quality_check_service.py"
      ],
      "suitable_metrics": [
        "precision",
        "recall",
        "F1",
        "IoU",
        "topology",
        "coverage"
      ],
      "task": "vector_comparison"
    },
    {
      "capability_id": "vector_clip_buffer_intersect",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "covered_by_family_not_separately_benchmarked",
      "implementation_kind": "deterministic_vector_processing",
      "implementation_paths": [
        "backend/app/services/vector_operations_service.py",
        "backend/app/services/vector_feature_service.py"
      ],
      "suitable_metrics": [
        "geometry validity",
        "CRS correctness",
        "area conservation",
        "feature counts",
        "topology"
      ],
      "task": "vector_comparison"
    },
    {
      "capability_id": "temporal_vector_change",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_contract_case_only",
      "implementation_kind": "deterministic_change_detection",
      "implementation_paths": [
        "backend/app/services/change_detection_service.py"
      ],
      "suitable_metrics": [
        "event precision",
        "event recall",
        "event F1",
        "IoU"
      ],
      "task": "change_detection"
    },
    {
      "capability_id": "thematic_raster_interpretation",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_metric_contract_only_no_generic_learned_classifier_claim",
      "implementation_kind": "deterministic_source_interpretation",
      "implementation_paths": [
        "backend/app/services/thematic_raster_analysis_service.py"
      ],
      "suitable_metrics": [
        "pixel accuracy",
        "per-class F1",
        "per-class IoU",
        "mean IoU"
      ],
      "task": "raster_classification"
    },
    {
      "capability_id": "raster_clip_reproject_indices",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "covered_by_family_not_separately_benchmarked",
      "implementation_kind": "deterministic_raster_processing",
      "implementation_paths": [
        "backend/app/services/raster_service.py",
        "backend/app/services/raster_operations_service.py"
      ],
      "suitable_metrics": [
        "CRS/transform preservation",
        "pixel alignment",
        "nodata",
        "numeric tolerance"
      ],
      "task": "raster_classification"
    },
    {
      "capability_id": "raster_partition_mosaic",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "not_separately_benchmarked",
      "implementation_kind": "deterministic_raster_partitioning",
      "implementation_paths": [
        "backend/app/services/raster_partition_analysis_service.py"
      ],
      "suitable_metrics": [
        "seam equality",
        "coverage completeness",
        "resolution consistency"
      ],
      "task": "raster_classification"
    },
    {
      "capability_id": "terrain_height_interpretation",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_contract_case_only",
      "implementation_kind": "deterministic_continuous_raster_analysis",
      "implementation_paths": [
        "backend/app/services/terrain_analysis_service.py",
        "backend/app/services/spw_terrain_service.py"
      ],
      "suitable_metrics": [
        "MAE",
        "RMSE",
        "bias",
        "coverage",
        "unit integrity"
      ],
      "task": "terrain_interpretation"
    },
    {
      "capability_id": "flood_hazard_interpretation",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "covered_by_family_not_separately_benchmarked",
      "implementation_kind": "deterministic_scenario_raster_analysis",
      "implementation_paths": [
        "backend/app/services/flood_hazard_analysis_service.py"
      ],
      "suitable_metrics": [
        "depth MAE/RMSE",
        "hazard-class IoU",
        "coverage",
        "scenario identity"
      ],
      "task": "terrain_interpretation"
    },
    {
      "capability_id": "bathymetry_interpretation",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "covered_by_family_not_separately_benchmarked",
      "implementation_kind": "deterministic_vertical_reference_analysis",
      "implementation_paths": [
        "backend/app/services/bathymetry_raster_analysis_service.py",
        "backend/app/services/mdk_bathymetry_probe_service.py"
      ],
      "suitable_metrics": [
        "MAE",
        "RMSE",
        "bias",
        "coverage",
        "vertical-datum integrity"
      ],
      "task": "terrain_interpretation"
    },
    {
      "capability_id": "data_contract_validation_and_scan",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "synthetic_contract_case_only",
      "implementation_kind": "deterministic_validation",
      "implementation_paths": [
        "backend/app/services/data_contract_validation.py",
        "scripts/run_accuracy_phase3_full_data_scan.py"
      ],
      "suitable_metrics": [
        "anomaly precision",
        "anomaly recall",
        "anomaly F1",
        "critical misses"
      ],
      "task": "geospatial_data_validation"
    },
    {
      "capability_id": "aoi_partition_orchestration",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "not_separately_benchmarked",
      "implementation_kind": "deterministic_orchestration",
      "implementation_paths": [
        "backend/app/services/aoi_operation_service.py",
        "backend/app/services/aoi_operation_executor.py",
        "backend/app/services/aoi_operation_worker.py"
      ],
      "suitable_metrics": [
        "partition completeness",
        "overlap/gap",
        "idempotency",
        "resume correctness"
      ],
      "task": "geospatial_data_validation"
    },
    {
      "capability_id": "geo_assistant_orchestration",
      "claim_boundary": "A family mapping is not evidence that this capability has a separate representative product benchmark.",
      "evaluation_status": "no_independent_accuracy_score_underlying_tool_results_are_authoritative",
      "implementation_kind": "tool_orchestration_interface",
      "implementation_paths": [
        "backend/app/services/geo_assistant_service.py"
      ],
      "suitable_metrics": [
        "tool-selection correctness",
        "grounding",
        "unsupported-claim rate"
      ],
      "task": "geospatial_data_validation"
    }
  ]
}