Early Detection of Diabetic Retinopathy Based Artificial Intelligent Techniques

Abstract

The eye is impacted by several disorders, either directly or indirectly. As a result, eye exams are a crucial component of general healthcare. One of the effects of diabetes is diabetic retinopathy (DR), which affects the blood vessels that supply and nourish the retina and causes severe visual loss. One of the prevalent eye conditions and a consequence of diabetes that affects the eyes is diabetic retinopathy. The symptoms of diabetic retinopathy may be absent or minimal. It may eventually result in blindness. Therefore, seeing symptoms early could aid in preventing blindness. This paper aims to research automatic methods for detecting diabetic retinopathy and create a reliable system for doing so. A modified extracted feature for the automatic identification of DR in digital fundus pictures is presented. The properties of exudates, blood vessels, and microaneurysms—three elements of diabetic retinopathy—are reported utilizing a variety of image processing techniques. Back Propagation Neural Networks (BPNN) and Support Vector Machine (SVM) classifiers are used to categorize the phases. SVM, which has accuracy, sensitivity, and specificity of 96.5, 97.2, and 93.3 percent, respectively, is the model that performs the best overall. The eye is impacted by several disorders, either directly or indirectly. As a result, eye exams are a crucial component of general healthcare. One of the effects of diabetes is diabetic retinopathy (DR), which affects the blood vessels that supply and nourish the retina and causes severe visual loss. One of the prevalent eye conditions and a consequence of diabetes that affects the eyes is diabetic retinopathy. The symptoms of diabetic retinopathy may be absent or minimal. It may eventually result in blindness. Therefore, seeing symptoms early could aid in preventing blindness. This paper aims to research automatic methods for detecting diabetic retinopathy and create a reliable system for doing so. A modified extracted feature for the automatic identification of DR in digital fundus pictures is presented. The properties of exudates, blood vessels, and microaneurysms—three elements of diabetic retinopathy—are reported utilizing a variety of image processing techniques. Back Propagation Neural Networks and Support Vector Machine classifiers are used to categorize the phases. SVM, which has accuracy, sensitivity, and specificity of 96.5, 97.2, and 93.3 percent, respectively, is the model that performs the best overall..