[공지사항]을 빙자한 안부와 근황 
Show more

Daily Arxiv

This is a page that curates AI-related papers published worldwide.
All content here is summarized using Google Gemini and operated on a non-profit basis.
Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.

BleedOrigin: Dynamic Bleeding Source Localization in Endoscopic Submucosal Dissection via Dual-Stage Detection and Tracking

Created by
  • Haebom

Author

Mengya Xu, Rulin Zhou, An Wang, Chaoyang Lyu, Zhen Li, Ning Zhong, Hongliang Ren

Outline

To address the challenges of developing an AI-based system for accurate, real-time localization and continuous monitoring of bleeding during endoscopic submucosal dissection (ESD), we present BleedOrigin-Bench, an ESD bleeding source dataset, and BleedOrigin-Net, a bleeding source localization and tracking framework. BleedOrigin-Bench contains 106,222 frames from 44 surgeries, 1,771 expert-annotated bleeding sources, covering eight anatomical sites and six clinical scenarios. BleedOrigin-Net is a dual-step detection-tracking framework that handles the entire workflow from bleeding event detection to continuous spatial tracking, and achieves state-of-the-art performance compared to existing object detection models and point tracking methods.

Takeaways, Limitations

•
Takeaways:
◦
Contribute to the advancement of AI-based bleeding detection and tracking research by first releasing the ESD bleeding origin dataset BleedOrigin-Bench.
◦
We present a novel framework, BleedOrigin-Net, for bleeding source location detection and tracking and verify its excellent performance.
◦
Suggesting the possibility of contributing to improving the efficiency of ESD surgery and ensuring patient safety through real-time bleeding detection and tracking.
•
Limitations:
◦
The dataset may be limited in size (a larger dataset with a wider range of patients, institutions, bleeding types, etc. may be needed).
◦
Additional validation of the generalization performance of BleedOrigin-Net is needed (performance evaluation in various endoscopic equipment and surgical environments)
◦
Additional research is needed for application in real clinical settings (evaluating the usability and clinical usefulness of the system).
👍
Made with Slashpage