Development Of Intraoperative Guidance Tools For Phacoemulsification Cataract Surgery

Cataract is the most common practiced surgery in the world, and the use of intraoperative microscopes during cataract surgery is a widespread practice worldwide. Taking advantage of these intraoperative systems, it is possible to post-process the images acquired in real time, providing meaningful feedback to the surgeon in the operating room. The goal of this thesis is to develop and evaluate a novel surgical guidance to help phacoemulsification cataract surgeons, residents, and medical students to improve their surgical outcomes and obtain better surgical training. Deep learning and computer vision techniques were studied to be applied in the operating room environment, with real-time data acquisition from the microscope camera video output. Different surgical guidance tools were developed for solving common problems that surgeons face in the operating room, and a final assessment survey was made by invited experts from the UIC Department of Ophthalmology & Visual Sciences to evaluate its usefulness and performance. Results suggest that deep learning real-time pupil tracking and phase identification during phacoemulsification cataract surgery is feasible, while computer vision-based tools are able to provide meaningful intraoperative feedback to surgeons and improve their overall experience.