Join us for the DCMB Tools & Technology Seminar Series featuring a presentation by Binh Duong Giap from the Department of Ophthalmology & Visual Sciences. Enjoy a complimentary pizza lunch while learning about innovative research and cutting-edge tools.

“Advancing Cataract Surgery Analytics with AI: Pupil Segmentation and Surgical Skill Metrics”

Abstract:
The growing availability of surgical video data creates opportunities to develop computational tools for automated analysis of surgical procedures. In cataract surgery, quantitative analysis of intraoperative videos can support research, surgical training, and objective assessment of surgical performance. However, extracting reliable visual features from surgical videos remains challenging due to instrument occlusion, lighting variations, and complex surgical motion.

In this seminar, I will present a computational framework for automated analysis of cataract surgery videos. First, I introduce tensor-based feature extraction methods for robust pupil recognition and segmentation during surgery. These methods enable reliable detection of pupil structures under challenging intraoperative conditions. I will also present CatSkill, an AI-based system that analyzes surgical videos to derive quantitative metrics for assessing surgical skill. Together, these approaches demonstrate how computer vision and machine learning can enable objective analysis of surgical videos and support the development of data-driven tools for ophthalmic surgery.