Uncertainty-Aware Domain Adaptation for Vitiligo Segmentation in Clinical Photographs

TL;DR

This paper introduces a frequency-aware, uncertainty-estimating segmentation framework for vitiligo in clinical photos, combining domain adaptation, architectural enhancements, and trust mechanisms to improve accuracy and reliability.

Contribution

It presents a novel, comprehensive framework integrating domain-adaptive training, a ConvNeXt V2-based encoder with spectral gating, and a trust mechanism with uncertainty maps for vitiligo segmentation.

Findings

Achieved a Dice score of 85.05% on clinical data.

Reduced boundary error significantly from 44.79 px to 29.95 px.

Demonstrated zero catastrophic failures and high reliability.

Abstract

Accurately quantifying vitiligo extent in routine clinical photographs is crucial for longitudinal monitoring of treatment response. We propose a trustworthy, frequency-aware segmentation framework built on three synergistic pillars: (1) a data-efficient training strategy combining domain-adaptive pre-training on the ISIC 2019 dataset with an ROI-constrained dual-task loss to suppress background noise; (2) an architectural refinement via a ConvNeXt V2-based encoder enhanced with a novel High-Frequency Spectral Gating (HFSG) module and stem-skip connections to capture subtle textures; and (3) a clinical trust mechanism employing K-fold ensemble and Test-Time Augmentation (TTA) to generate pixel-wise uncertainty maps. Extensive validation on an expert-annotated clinical cohort demonstrates superior performance, achieving a Dice score of 85.05% and significantly reducing boundary error (95% Hausdorff Distance improved from 44.79 px to 29.95 px), consistently outperforming strong CNN (ResNet-50 and UNet++) and Transformer (MiT-B5) baselines. Notably, our framework demonstrates high reliability with zero catastrophic failures and provides interpretable entropy maps to identify ambiguous regions for clinician review. Our approach suggests that the proposed framework establishes a robust and reliable standard for automated vitiligo assessment.