Diabetes Retinopathy Detection

Diabetic Retinopathy (DR) results from damage to the retina, potentially leading to vision loss. DR often lacks early warning signs, gradually promoting the formation of new blood vessels at the back of the eye. This can lead to blood clots, bleeding, and distorted vision. Traditional methods have not achieved optimal classification accuracy. Consequently, this article focuses on the implementation of a hybrid logistic regression (HLR) based machine learning model for classifying DR. The process begins with histogram equalization to enhance the DR image region. Next, the segmentation of microaneurysms is conducted using image morphological operations. Features are then extracted using the gray level co-occurrence matrix (GLCM), which reveals the internal relationships associated with DR.The selection of features is conducted using the Gaussian Mixture Model (GMM). Subsequently, the HLR model is employed to execute the multi-class classification task. Simulation results indicate that the proposed method outperforms existing state-of-the-art approaches.