Abstract :

Diabetic Retinopathy (DR) is a leading cause of blindness among diabetic patients, necessitating early detection for effective treatment. In the existing systems of Machine Learning (ML) approaches like Support Vector Machine (SVM) is employed for DR classification using handcrafted features such as Blood vessels overcomes the drawbacks of SVM and offers the improved classification, accuracy and better generalization for large-scale datasets. This concludes that contribution to early DR detection and improved patient’s care at remote areas also extracted from preprocessed retinal images due to this the SVM is struggled with non-linear patterns in medical imaging. To overcome these challenges, the novelty of proposed Deep Learning (DL) algorithm such as Deep Neural Network (DNN) model which automatically learns hierarchical features, enabling more accuracy and better classification. The performance of both models is evaluated by using the performance Metrics like Accuracy, Precision a