modosaic.segmentation.validators.impl.mask_statistics¶

modosaic.segmentation.validators.impl.mask_statistics ¶

MaskStatsValidator ¶

Bases: SegmentationValidator[MaskValidationStats]

Compute coverage, distinctness, and fragmentation scores for masks.

validate ¶

validate(record, generated)

Validate generated masks with aggregate mask statistics.

Source code in modosaic/segmentation/validators/impl/mask_statistics.py

@override
def validate(self, record: ImageRecord, generated: list[np.ndarray]) -> MaskValidationStats:
    """Validate generated masks with aggregate mask statistics."""
    logger.debug(f"Computing mask statistics for record sample {record.sample_id} with {len(generated)} masks")
    stats = self._compute_score(generated)
    logger.debug(f"Computed mask statistics for record sample {record.sample_id}")
    return stats