A Deep Learning Model for Five Stage Diabetic Retinopathy Detection

Diabetic retinopathy (DR) is a common complication of diabetes that can result in vision loss if not diagnosed and treated in a timely manner. DR has several risk factors, including the duration of diabetes, poor control of blood sugar, hypertension, and high cholesterol levels. It is important for diabetics to undergo regular eye examinations in order to detect and manage this condition as early as possible. In order to slow down the progression of DR, healthcare professionals must be able to detect and classify the disease early in order to implement timely interventions and treatment strategies. Thus, it is possible to prevent the development of further visual impairments. This paper presents a novel approach for detecting and classifying DR stages based on image data. An array of retinal images obtained from diabetic patients was analyzed using state-of-the-art deep learning techniques. As a result, it aims to automate the detection and classification of Diabetic Retinopathy across its five stages. A pre-trained EfficientNetB3 model was applied with the transfer learning technique and achieved 90.03 percent accuracy. By employing this approach, this debilitating disease could be diagnosed and treated more effectively.