One of the principal causes of blindness is Diabetic Retinopathy (DR), a medical condition in which damage occurs to the retina due to diabetes. According to the United States National Library of Diabetes, more than 40 percent of Americans diagnosed with diabetes have some stage of Diabetic Retinopathy. The International Diabetes Federation estimates that are 415 million diabetic patients worldwide have the risk of DR.
Ophthalmologists diagnose the diabetic eye disease by examining the pictures of the back of the eye for lesions. The severity is identified by the presence of lesions that indicate bleeding and fluid leakage in the eye. Only experienced doctors and specialists can scan and interpret these pictures. There is an acute shortage of qualified ophthalmologists that can diagnose diabetic eye, which puts many patients at the risk of losing their eyesight.
Google Brain, the research team within Google that focuses on the application of AI, has collaborated with doctors in India and US to help them diagnose DR. The team has collected over 128,000 images that were each evaluated by 3-7 ophthalmologists from a panel of 54 ophthalmologists. These images were fed to a deep learning algorithm for creating a model to detect Diabetic Retinopathy. The performance of the algorithm was tested on two different datasets totaling to 12,000 images.