Predicting Diabetic Retinopathy Using Deep Learning

Abstract

Abstract: Research works done so far are majorly focused on the risk factors which include diabetic retinopathy (DR); however, it is still unclear up to what extent the risk factors are associated with diabetic retinopathy. We use early prevention procedures considering diabetic retinopathy during the most high-risk group with better detected DR-related risk factors. Machine learning, a recent advancement during state-of-the-art technology, plays a critical and crucial role during image processing applications such as biomedical, satellite image processing & Artificial Intelligence applications includes object identification & recognition, among others. The goal of this study is to look towards a deep-learning system to predict the probability of diabetic retinopathy developing within two years among people with diabetes. Using colour fundus pictures, deep-learning algorithms predict the development regarding diabetic retinopathy and the diabetic systems were independent which were more informative to the existing risk variables. The proposed works aims to develop and validate the deep learning system towards the prediction of progression regarding diabetic retinopathy of diabetic individuals to receive tele-retinal diabetic retinopathy screening during a primary care environment.

Keywords: Diabetic Retinopathy, Deep learning, Color fundus photographs, Screening.