A decision-aware evaluation protocol is proposed 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.
Abstract
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. The protocol reports fixed-budget error reduction, risk-coverage behaviour, region-level review proxies, structural failure modes, probability-derived baselines, and sensitivity to review assumptions. We instantiate it across retinal vessel, skin lesion, and polyp segmentation with multiple backbones and uncertainty estimators. 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. However, estimator rankings are not universal. For example, probability-derived scores remain competitive, deep ensembles can outperform TTA, learned deferral helps mainly when the base estimator is weak, and rankings can change with review unit, uncertainty target, correction assumptions, and cost model. Structural and second-observer analyses further show that boundary and thin-structure failures are partly tied to annotation ambiguity. These findings support evaluating uncertainty as an estimator-policy-review configuration, not as a standalone score map. Code: https://anonymous.4open.science/r/Uncertainity-in-Medical-Image-Segmentation/README.md
Segmentation masks are increasingly reviewed rather than redrawn, which requires knowing where a mask is likely wrong. Learned error maps rank pixels by risk, but the number of pixels to warn is chosen ad hoc, and calibrating one warning threshold over all error pixels may leave errors far from the predicted contour un...
Nishi Kanta Paul, Md Minhajul Islam Shovo, Hoi Leong Lee· 0 citations
Reliable medical-image classification requires assessment of accuracy, calibration, and uncertainty under image degradation. We analyze the unit-prior Evidential Mean Squared Error (EMSE) objective with masked Kullback–Leibler (KL) regularization and evaluate four ResNet-18 classifiers for four-class blood-smear classi...
Umar Hasan, M. Martuza, Turki G. Alghamdi· Italian National Conference...· 0 citations
Modern Deep-Learning facilitates automatic segmentation in large imaging studies. To prevent segmentation errors from impacting data analysis, automated quality control (QC) for filtering segmentations and identifying errors (segmentation selection) is essential. QC must be fast and require minimal expert delineations...
T. Norajitra, T. Nonnenmacher, T. Haueise et al.· Medical Image Analysis· 1 citation
Reliability under sparse and heterogeneous failures remains a fundamental challenge for medical image segmentation. High average accuracy can conceal a small set of structurally distinct and clinically consequential errors. Existing post-hoc correction methods alleviate this problem, but typically estimate false-positi...
Zi-Liang Wang, Xu-Jiang Tang, Yu-Ting Lu et al.· arXiv.org· 0 citations
PURPOSE
Despite remarkable auto-segmentation performance and growing clinical adoption, models frequently fail to generalize beyond the training distribution. Therefore, this study aimed to develop and validate an uncertainty-aware segmentation pipeline that yields high delineation accuracy for liver cancer patients an...
D. Langner, C. Gani, S. Boeke et al.· Radiotherapy and Oncology· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.