Towards Reliable WMH Segmentation under Domain Shift: An Application Study using Maximum Entropy Regularization to Improve Uncertainty Estimation

2025-06-17

AI Summary

To address domain shift in white matter hyperintensity (WMH) segmentation across multi-center MRI due to variations in scanners and acquisition protocols, this paper proposes a calibration-augmented method for deployment-time unsupervised error identification. The core innovation is the first integration of maximum entropy regularization into the WMH segmentation framework: it jointly optimizes segmentation performance with Dice loss within a U-Net architecture while leveraging entropy estimation to quantify prediction uncertainty—thereby improving its correlation with actual segmentation errors. Critically, the method operates without ground-truth labels, enabling robust cross-scanner calibration and error detection. Experiments demonstrate a 32% reduction in expected calibration error (ECE), a 0.41 increase in Spearman’s ρ between uncertainty and error, and stable Dice scores—collectively enhancing model reliability and clinical applicability on out-of-distribution data.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Segmentation
• Reasoning under Uncertainty: Uncertainty Representations

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Accurate segmentation of white matter hyperintensities (WMH) is crucial for clinical decision-making, particularly in the context of multiple sclerosis. However, domain shifts, such as variations in MRI machine types or acquisition parameters, pose significant challenges to model calibration and uncertainty estimation. This study investigates the impact of domain shift on WMH segmentation by proposing maximum-entropy regularization techniques to enhance model calibration and uncertainty estimation, with the purpose of identifying errors post-deployment using predictive uncertainty as a proxy measure that does not require ground-truth labels. To do this, we conducted experiments using a U-Net architecture to evaluate these regularization schemes on two publicly available datasets, assessing performance with the Dice coefficient, expected calibration error, and entropy-based uncertainty estimates. Our results show that entropy-based uncertainty estimates can anticipate segmentation errors, and that maximum-entropy regularization further strengthens the correlation between uncertainty and segmentation performance while also improving model calibration under domain shift.

Problem

• Improves WMH segmentation accuracy under domain shifts
• Enhances uncertainty estimation without ground-truth labels
• Strengthens correlation between uncertainty and segmentation performance

Innovation

• Maximum-entropy regularization enhances uncertainty estimation
• U-Net architecture evaluates domain shift impact
• Entropy-based uncertainty predicts segmentation errors