Deep Learning Approaches for MultiDisease Retinopathy Detection and Classification

Retinal diseases such as Diabetic Retinopathy (DR), Age-related Macular Degeneration (AMD), Glaucoma, and Cataract are major causes of vision impairment and blindness worldwide. Early detection and classification of these conditions are critical for effective treatment and prevention of irreversible damage. This project presents a deep learning-based system for multi-disease retinopathy detection and classification using retinal fundus images. The dataset, sourced from Kaggle, includes images categorized into six classes: Dry AMD, Mild DR, Glaucoma, Wet AMD, Cataract, and Normal. Two deep learning models—Convolutional Neural Networks (CNN) and MobileNet—are implemented to achieve efficient and accurate classification. The front-end of the system is developed using HTML, CSS, and JavaScript, providing modules for user registration, login, and image-based prediction. The back-end is built on Flask with Python, enabling secure image upload, preprocessing, and prediction. This system aims to support ophthalmologists and healthcare professionals in automated diagnosis and decision-making.

Keywords:

Deep Learning, Retinopathy Detection, CNN, MobileNet, Fundus Images, Diabetic Retinopathy, Glaucoma, AMD, Cataract, Flask.