Mix, Align, Distil: Reliable Cross-Domain Atypical Mitosis Classification
This work addresses domain-robust classification for histopathological markers, which is crucial for medical imaging but incremental in method.
The paper tackled the problem of consistently identifying atypical mitotic figures under domain shift from scanner, stain, and acquisition differences, achieving a balanced accuracy of 0.8762 and ROC AUC of 0.9499 on the MIDOG 2025 Task 2 leaderboard.
Atypical mitotic figures (AMFs) are important histopathological markers yet remain challenging to identify consistently, particularly under domain shift stemming from scanner, stain, and acquisition differences. We present a simple training-time recipe for domain-robust AMF classification in MIDOG 2025 Task 2. The approach (i) increases feature diversity via style perturbations inserted at early and mid backbone stages, (ii) aligns attention-refined features across sites using weak domain labels (Scanner, Origin, Species, Tumor) through an auxiliary alignment loss, and (iii) stabilizes predictions by distilling from an exponential moving average (EMA) teacher with temperature-scaled KL divergence. On the organizer-run preliminary leaderboard for atypical mitosis classification, our submission attains balanced accuracy of 0.8762, sensitivity of 0.8873, specificity of 0.8651, and ROC AUC of 0.9499. The method incurs negligible inference-time overhead, relies only on coarse domain metadata, and delivers strong, balanced performance, positioning it as a competitive submission for the MIDOG 2025 challenge.