Expeditious Detection of Diabetic Retinopathy Using ResNet 50 CNN Algorithm

If identified and treated promptly, diabetic retinopathy, a severe complication of diabetes, can cause blindness. The early detection of diabetic retinopathy requires automated devices because manual detection by ophthalmologists is time-consuming. This paper suggests a machine-learning method for detecting diabetic retinopathy based on fundus pictures. Before being classified, characteristics like exudates, hemorrhages, and microaneurysms are extracted using a support vector machine. The accuracy of tests has reached up to 82%, showing the success of this method at identifying diabetic retinopathy. This research supports ongoing initiatives to create automatic tools for the early diagnosis of retinopathy, which is crucial for the prompt treatment and verging blindness in diabetic patients.