Confdence Contours: Uncertainty-Aware Annotation for Medical Semantic Segmentation

Abstract

Medical image segmentation modeling is a high-stakes task where understanding of uncertainty is crucial for addressing visual ambiguity. Prior work has developed segmentation models utilizing probabilistic or generative mechanisms to infer uncertainty from labels where annotators draw a singular boundary. However, as these annotations cannot represent an individual annotator’s uncertainty, models trained on them produce uncertainty maps that are diffcult to interpret. We propose a novel segmentation representation, Confdence Contours, which uses high- and low-confdence “contours” to capture uncertainty directly, and develop a novel annotation system for collecting contours. We conduct an evaluation on the Lung Image Dataset Consortium (LIDC) and a synthetic dataset. From an annotation study with 30 participants, results show that Confdence Contours provide high representative capacity without considerably higher annotator effort. We also fnd that general-purpose segmentation models can learn Confdence Contours at the same performance level as standard singular annotations. Finally, from interviews with 5 medical experts, we fnd that Confdence Contour maps are more interpretable than Bayesian maps due to representation of structural uncertainty