AI RESEARCH

CLEAR: Causal Learning Framework For Robust Histopathology Tumor Detection Under Out-Of-Distribution Shifts

arXiv CS.CV

ArXi:2510.14273v2 Announce Type: replace Domain shift in histopathology, often caused by differences in acquisition processes or data sources, poses a major challenge to the generalization ability of deep learning models. Existing methods primarily rely on modeling statistical correlations by aligning feature distributions or