A Deep Learning Approach to Analyze Diabetic Retinopathy Lesions using Scant Data

One of the most dangerous effects of diabetes is diabetic retinopathy, which, if ignored, results in lifelong blindness. Early detection, which is crucial for successful treatment outcomes, is one of the key obstacles. Unfortunately, it takes a skilled human to accurately evaluate fundus images to determine the exact diabetic retinopathy stage. Millions of people can benefit from the detecting step's simplification. The identification of diabetic retinopathy is one area where convolutional neural networks (CNN) have been successfully utilized. Early detection may prevent the possibility of permanent and complete blindness. Therefore, an efficient screening system is needed. We proposed a system that can classify various stages of diabetic retinopathy to facilitate the screening process.