CFR-Net:Collaborative Feature Refinement Network for Medical Image Anomaly Detection
Medical image anomaly detection is central to timely diagnosis and clinical decision support, yet abnormal samples are costly to collect because of disease rarity, privacy concerns, and expert workload. This motivates unsupervised learning from normal images, where abnormalities are detected as deviations from learned normal patterns. However, medical anomalies are often subtle, local, and intertwined with normal anatomical variations, which complicates reliable normality modeling. Distillation-based methods support normality modeling by using frozen pretrained teachers as stable feature references, yet mismatches between generic teacher priors and student representations adapted to medical images can produce residuals unrelated to abnormalities in conventional distillation pipelines. To address this limitation, we propose the Collaborative Feature Refinement Network, which learns normality through a coupled process of shared feature conditioning before decoding and cross-space consistency after decoding. Shared feature conditioning performs medical-aware conditioning on teacher and student features under common rules, while cross-space consistency constrains each decoded stream with the complementary encoder representation for reciprocal normal reconstruction. The coupled process is further stabilized by the homework set reorganization strategy, which periodically refreshes normal training subsets. Experiments on six medical image benchmarks show competitive anomaly classification and strong anomaly localization performance.