A Decision-Aware Evaluation Protocol for Uncertainty in Medical Image Segmentation

Key points

• We propose a decision-aware evaluation protocol that separates uncertainty estimation from deferral policy and evaluates their combination under fixed review budgets: which pixels, regions, or images should be accepted, and which should be reviewed.1
• We instantiate it across retinal vessel, skin lesion, and polyp segmentation with multiple backbones and uncertainty estimators.1
• Under matched inference budgets and canonical multi-layer MC Dropout, test-time augmentation (TTA) gives a stronger error-detection signal across matched MC--TTA comparisons, including architecture and dropout-placement checks.1
• Structural and second-observer analyses further show that boundary and thin-structure failures are partly tied to annotation ambiguity.1

Sources (1)

• [1]A Decision-Aware Evaluation Protocol for Uncertainty in Medical Image Segmentation
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 09:59 AM

  Medical segmentation uncertainty is often evaluated with calibration error or pixel-level error-detection scores, but these metrics do not show whether uncertainty improves downstream review decisions.

  We propose a decision-aware evaluation protocol that separates uncertainty estimation from deferral policy and evaluates their combination under fixed review budgets: which pixels, regions, or images should be accepted, and which should be reviewed.

Extractive summary: sentences quoted from the sources.