RDX is an innovative artificial intelligence-powered system designed for the automated detection and staging of sight-threatening retinal diseases through deep learning analysis of fundus images. The system addresses a critical healthcare challenge by enabling early identification of three major retinal conditions, Retinoblastoma, Diabetic Retinopathy, and Age-Related Macular Degeneration, across two progressive stages each. Leveraging Python for backend logic, OpenCV for image preprocessing and feature enhancement, and TensorFlow for implementing convolutional neural networks, the system learns to recognize disease-specific biomarkers including microaneurysms, hard exudates, drusen deposits, and neovascularization. The architecture encompasses an upload module supporting common image formats with comprehensive validation and a camera capture module enabling real-time fundus photography through Android devices. Upon analysis, the CNN-based classification engine generates predictions with confidence scores, while an integrated disease database provides diagnostic summaries including symptomatology, etiology, stage-specific treatment protocols, severity assessment, and urgency levels.

Keywords— Retinal disease detection; Deep learning; Convolutional neural network; Diabetic retinopathy; Retinoblastoma; Age-related macular degeneration; Fundus image analysis