Abstract
Ophthalmology is heavily dependent on imaging and numerical data, making it an excellent candidate for the application of deep learning to tasks in image analysis and clinical decision support. In this seminar, we will discuss the rapid automated segmentation of anterior segment optical coherence tomography (AS-OCT) and its implications for the clinical investigation of the cornea and the intraoperative guidance of surgical maneuvers. In addition, we will discuss the relevance of deep learning to lens implant selection for cataract surgery -- the most commonly performed surgical procedure in the United States. We will conclude by examining the potential roles for deep learning in the analysis of the SOURCE database -- a comprehensive repository of ophthalmic clinical and imaging data being built at UM Kellogg Eye Center to encompass data across 18 institutions.