RED-Sphere: Hyperspherical Residual Edge Debiasing for Cross-Population Fundus Disease Domain Generalization
2026-07-12 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address the problem that medical image classifiers trained on one group of patients may not work well on others with different features or disease patterns. They propose a method called RED-Sphere that helps the classifier focus on important disease features rather than misleading image cues that vary between populations. Their approach works without using any information from external populations during training or tuning. Tested on eye images for two diseases, RED-Sphere improved classification performance on unseen patient groups. The method can be adapted to other medical imaging tasks where harmless shortcuts and real disease signals are mixed.
Medical Image ClassificationPopulation ShiftRobustnessShortcut LearningSpherical PrototypesCounterfactual ConsistencyScanning Laser OphthalmoscopyAge-Related Macular DegenerationDiabetic RetinopathyDomain Generalization
Authors
Yan Lin, Ziheng Wang, Shuang Chen, Amir Atapour-Abarghouei, Stephen McGough
Abstract
Medical image classifiers are often trained within one source population, yet clinical deployment requires robustness to patients whose appearance, acquisition style, and disease prevalence differ from the source cohort. Existing fairness and robustness methods often require group supervision or treat appearance variation as an undifferentiated nuisance, which is insufficient when population-correlated low-level cues and lesion evidence share edge and texture structure. We study a strict source-only cross-population setting, where external populations are unseen during optimization, validation, scheduling, hyperparameter and model selection. We propose RED-Sphere, a plug-and-play robustness framework for image classification under unseen population shifts. It estimates shortcut-sensitive nuisance responses with an edge and feature energy prior, attenuates dominant responses through residual soft gating, regularizes masked nuisance views with counterfactual-inspired consistency and separation losses, and predicts labels with normalized spherical prototypes. It favours angular semantic evidence over source-correlated activation magnitude while preserving lesion structure. Although demonstrated on 2D Scanning Laser Ophthalmoscopy (SLO) fundus classification for Age-Related Macular Degeneration (AMD) and Diabetic Retinopathy (DR), RED-Sphere is not tied to retinal anatomy: the same principle can be adapted with modality-specific nuisance priors wherever appearance shortcuts and semantic evidence are entangled. Under a strict White-only Harvard-FairVision protocol, RED-Sphere improves held-out macro-F1 across all 20 task and backbone comparisons, with average gains of 1.28 and 2.98 F1 points on AMD and DR. Gains in AUC and PR-AUC, visual diagnostics, ablations, and sensitivity analyses further support stronger external semantic alignment and more stable angular disease geometry.