An AI-powered tool leveraging ResNet-50 and Xception pretrained models to detect diabetic retinopathy with 97.5% accuracy. Upload a retinal image to instantly classify severity levels, aiding in early diagnosis and vision loss prevention.