Overview
Endoscopy research studies how to interpret images and videos from procedures such as colonoscopy, capsule endoscopy, bronchoscopy, and laryngoscopy, as well as how to guide robotic instruments. Common approaches include neural networks and vision-language models for detection, classification, segmentation, and reporting; temporal models for analyzing long videos; and depth estimation, mapping, and reinforcement learning for navigation. Datasets and benchmarks support evaluation of tasks ranging from polyp detection and disease scoring to 3D reconstruction and autonomous control. A central challenge is making systems reliable across different devices, patients, and real-world procedures, while ensuring that their visual evidence and actions are clinically sound.
Latest papers
87 papers5 papers in the last four weeks, down 58% on the four weeks before. 0.0% of all new papers.