Geometrically-correct, DICOM-native evaluation metrics for medical image segmentation in Python: Dice, IoU, HD95, ASSD, Normalized Surface Dice, per-lesion detection F1, with bootstrap confidence intervals and subgroup/fairness slicing.
python dice metrics evaluation dicom medical-imaging segmentation image-segmentation nifti radiology medical-image-analysis hausdorff-distance rtstruct dicom-seg monai nnunet segmentation-metrics hd95 surface-distance surface-dice
-
Updated
Jun 20, 2026 - Python