Weakly Supervised Object Detection in Chest X-Rays with Differentiable ROI Proposal Networks and Soft ROI Pooling

arXiv:2402.11985v1 Announce Type: cross
Abstract: Weakly supervised object detection (WSup-OD) increases the usefulness and interpretability of image classification algorithms without requiring additional supervision. The successes of multiple instance learning in this task for natural images, however, do not translate well to medical images due to the very different characteristics of their objects (i.e. pathologies). In this work, we propose Weakly Supervised ROI Proposal Networks (WSRPN), a new method for generating bounding box proposals on the fly using a specialized region …

arxiv cs.cv cs.lg detection differentiable networks pooling roi type