Intelligent Diagnosis of Diabetic Retinal Disease Using Image Processing and Machine Learning Algorithms

Abstract

Untreated diabetic retinopathy, which develops as a result of prolonged elevated blood glucose levels, can ultimately cause permanent vision loss if it is not detected and managed at an early stage. Therefore, timely diagnosis and appropriate medical intervention are essential to prevent severe complications associated with this condition. However, manual diagnosis of diabetic retinopathy is often challenging and time-consuming, leading to delays in patients receiving consultation and treatment from ophthalmologists. To address this issue, automated diagnostic systems can assist in the early detection of diabetic retinopathy, enabling prompt treatment and reducing the risk of further ocular damage. The present study proposes a machine learning–based approach for extracting key retinal features, including exudates, hemorrhages, and micro aneurysms, and classifying them using a hybrid classifier that integrates Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Random Forest, Logistic Regression, and Multilayer Perceptron (MLP) models.