Uncertainty-Supervised Interpretable and Robust Evidential Segmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Yuzhu, Sui, An, Wu, Fuping, Zhuang, Xiahai
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917010589351936
author Li, Yuzhu
Sui, An
Wu, Fuping
Zhuang, Xiahai
author_facet Li, Yuzhu
Sui, An
Wu, Fuping
Zhuang, Xiahai
contents Uncertainty estimation has been widely studied in medical image segmentation as a tool to provide reliability, particularly in deep learning approaches. However, previous methods generally lack effective supervision in uncertainty estimation, leading to low interpretability and robustness of the predictions. In this work, we propose a self-supervised approach to guide the learning of uncertainty. Specifically, we introduce three principles about the relationships between the uncertainty and the image gradients around boundaries and noise. Based on these principles, two uncertainty supervision losses are designed. These losses enhance the alignment between model predictions and human interpretation. Accordingly, we introduce novel quantitative metrics for evaluating the interpretability and robustness of uncertainty. Experimental results demonstrate that compared to state-of-the-art approaches, the proposed method can achieve competitive segmentation performance and superior results in out-of-distribution (OOD) scenarios while significantly improving the interpretability and robustness of uncertainty estimation. Code is available via https://github.com/suiannaius/SURE.
format Preprint
id arxiv_https___arxiv_org_abs_2509_17098
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Uncertainty-Supervised Interpretable and Robust Evidential Segmentation
Li, Yuzhu
Sui, An
Wu, Fuping
Zhuang, Xiahai
Computer Vision and Pattern Recognition
Machine Learning
Uncertainty estimation has been widely studied in medical image segmentation as a tool to provide reliability, particularly in deep learning approaches. However, previous methods generally lack effective supervision in uncertainty estimation, leading to low interpretability and robustness of the predictions. In this work, we propose a self-supervised approach to guide the learning of uncertainty. Specifically, we introduce three principles about the relationships between the uncertainty and the image gradients around boundaries and noise. Based on these principles, two uncertainty supervision losses are designed. These losses enhance the alignment between model predictions and human interpretation. Accordingly, we introduce novel quantitative metrics for evaluating the interpretability and robustness of uncertainty. Experimental results demonstrate that compared to state-of-the-art approaches, the proposed method can achieve competitive segmentation performance and superior results in out-of-distribution (OOD) scenarios while significantly improving the interpretability and robustness of uncertainty estimation. Code is available via https://github.com/suiannaius/SURE.
title Uncertainty-Supervised Interpretable and Robust Evidential Segmentation
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2509.17098