Article’s

– Diabetic Retinopathy (DR) is a progressive eye disorder caused by diabetes and remains one of the major contributors to vision impairment worldwide. Early identification of retinal abnormalities is essential for preventing severe visual complications; however, conventional screening methods rely heavily on expert ophthalmologists and are often time-consuming. This study presents an intelligent Diabetic Retinopathy detection framework that utilizes deep learning techniques for automated analysis of retinal fundus images. The proposed approach incorporates image preprocessing, feature extraction, and a convolutional neural network-based classification model to categorize retinal images into five severity levels: No DR, Mild, Moderate, Severe, and Proliferative DR. To enhance model transparency, Gradient weighted Class Activation Mapping (Grad-CAM) is employed to highlight critical retinal regions influencing the prediction process. A web-based application is developed to provide a user-friendly platform for image upload, disease prediction, and result visualization. Experimental evaluation demonstrates the effectiveness of the proposed system in accurately identifying disease stages while supporting interpretable decision-making. The developed solution has the potential to assist healthcare professionals in large-scale screening programs, reduce diagnostic workload, and improve access to early retinal disease assessment. The study highlights the growing role of artificial intelligence in advancing medical image analysis and supporting efficient healthcare delivery.