Talk Abstract: Diabetic retinopathy is a major cause of preventable blindness worldwide. Diabetic retinopathy screening faces major challenges in terms of cost and accessibility, particularly for smaller and medium-sized practices serving large patient populations in need. Existing screening methods rely on expensive equipment, skilled personnel, and limited economies of scale, making them financially unsustainable for many clinics. AI-based grading does not fully address the challenge as it relies on high quality retina images that are difficult to capture in diabetic population. I describe advancements in two key areas related to retina imaging: new camera systems with on-device intelligence minimize operator skill dependence and significantly reduce the number of ungradable images. The utilization of high-speed imaging combined with powerful computational resources further unlocks new possibilities for capturing retinal photos, by addressing issues like cataracts and pupil size limitations. By focusing on cost-effective technologies, we can make diabetic retinopathy screening more accessible and affordable for underserved communities, ultimately leading to improved patient outcomes and a reduction in vision loss.

Speaker Biography: Sam Kavusi is the head of Retina Imaging and Teleretinal Integration at Verily. He is interested in improving availability and accuracy of retina image captures especially outside of the ophthalmic environments. Generally, he is interested in the application of on-device intelligence, computational imaging, artificial intelligence, and consumer electronics in the development of modern medical devices. He received the B.S. degree (Hons.) from Sharif University, Tehran, Iran in 1999, and the M.S. and PhD. degrees in electrical engineering from Stanford University, Stanford, CA in 2001 and 2006, respectively. He has held various industry positions leading and developing smartphone cameras, semiconductor/MEMS sensors and proteomic chips at Google and Bosch. He is co-inventor of more than 70 patents and his publications are cited more than 800 times. _________