PathNavigate: A Training-Free Pathology Agent with Surprise-Guided Scan and Shared Slide Memory for Whole-Slide Image VQA
Researchers have developed PathNavigate, a novel training-free agent designed for whole-slide image visual question answering in pathology. This agent employs a unique scan-search-readout routine, utilizing a surprise field to identify abnormal regions before focusing on question-conditioned targets. PathNavigate leverages a shared online memory and frozen pathology features to improve accuracy and interpretability in analyzing gigapixel pathology slides. AI
IMPACT Introduces a novel training-free approach for pathology image analysis, potentially improving diagnostic efficiency and accuracy.