Skip to content
#

hd95

Here are 2 public repositories matching this topic...

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.

  • Updated Jun 20, 2026
  • Python

Independent, contamination-aware, statistically-rigorous leaderboard for medical image segmentation models: every score with a bootstrap confidence interval and a ranking-stability test, sliced by failure mode, reproducible with one command. Powered by segauge.

  • Updated Jun 21, 2026
  • Python

Add this topic to your repo

To associate your repository with the hd95 topic, visit your repo's landing page and select "manage topics."

Learn more