CNN Based Model for Early Glaucoma Detection from Retinal Fundus Images

Glaucoma is a leading cause of irreversible blindness worldwide, and early detection is crucial for preventing vision loss. This project, titled "CNN-Based Glaucoma Detection Using Retinal Fundus Images", presents a deep learning framework for the accurate classification of glaucoma in retinal images. The system leverages Convolutional Neural Networks (CNN), MobileNet, and ResNet models for robust feature extraction and classification. The models are trained on the HIGAN-CNN Glaucoma Detection dataset from Kaggle, which contains labeled retinal fundus images for detecting glaucoma and normal cases. The image preprocessing pipeline involves resizing, normalization, and feature extraction using the pre-trained MobileNet and ResNet models, followed by classification via CNN-based architectures. The system is deployed through a Flask web interface, enabling users to upload retinal images and receive immediate glaucoma detection results. The modular architecture supports potential future expansions, including additional models and techniques for enhanced accuracy. This approach demonstrates the power of deep learning in medical image classification, offering an efficient and scalable solution for real-time glaucoma detection. Future work will focus on model optimization, real-time monitoring, and integration of advanced features for broader ophthalmic disorder detection.

Keywords
Glaucoma, CNN, MobileNet, ResNet, Retinal Fundus Images, Deep Learning, Flask Deployment, Medical Image Classification, Real-time Detection, Transfer Learning.