Retinal OCT Disease Classification

Deep learning system for classifying retinal diseases (CNV, DME, DRUSEN, Normal) from OCT scans using Custom CNN and VGG19 models.

Key Features

Clinical-grade performance: 92.13% test accuracy

Explainable AI: Grad-CAM visualization

️ Class balancing: Automatic weight calculation

️ Preprocessing pipeline: Denoising, CLAHE, ROI extraction

? Two model variants: Lightweight CNN (91.28%) and VGG19 (92.13%)