AI RESEARCH
A Lightweight Multi-Cancer Tumor Localization Framework for Deployable Digital Pathology
arXiv CS.AI
•
ArXi:2603.08844v1 Announce Type: cross Accurate localization of tumor regions from hematoxylin and eosin-stained whole-slide images is fundamental for translational research including spatial analysis, molecular profiling, and tissue architecture investigation. However, deep learning-based tumor detection trained within specific cancers may exhibit reduced robustness when applied across different tumor types. We investigated whether balanced