Publication

Abstract : Diabetic retinopathy is a chronic disease that affects millions of people worldwide. It is a disease that leads to vision loss in middle-aged people worldwide. Type 1 diabetes (T1D), type 2 diabetes (T2D), and gestational diabetes mellitus (GDM) are the three types of diabetes. Type 1 diabetes arises from the immune system's destruction of the pancreatic β-cells, which secrete insulin. Insufficient insulin production leads to type 2 diabetes. GDM happens while a woman is pregnant. Blood glucose levels in diabetics should be kept within a normal range. If not, it causes microvascular or macrovascular problems in the long run, as well as hyperglycemia or hypoglycemia in short-term difficulties. This article tackles the Indian Diabetic Retinopathy image (IDRiD) Dataset. Based on the degree of DR severity, the images are categorized into five classes: no retinoapathy (stage 0), mild (stage 1), moderate (stage 2), severe (stage 3) and proliferative (stage 4). Deep learning methods are commonly applied to diabetes in healthcare. The three categories in application of diabetes are diagnosis of diabetes, glucose management, and diagnosis of complications. Using deep learning techniques, we can provide solutions to DR problems.