Training Like a Medical Resident: Context-Prior Learning Toward Universal Medical Image Segmentation

arXiv:2306.02416v3 Announce Type: replace
Abstract: A major focus of clinical imaging workflow is disease diagnosis and management, leading to medical imaging datasets strongly tied to specific clinical objectives. This scenario has led to the prevailing practice of developing task-specific segmentation models, without gaining insights from widespread imaging cohorts. Inspired by the training program of medical radiology residents, we propose a shift towards universal medical image segmentation, a paradigm aiming to build medical image understanding foundation models by leveraging the diversity …

arxiv context cs.cv image medical prior segmentation training type universal