Ocular Disease Identification Using Deep Learning

Ocular diseases such as glaucoma, diabetic retinopathy, and age-related macular degeneration pose significant threats to vision and require early detection for effective management. This paper explores the application of deep learning-based convolutional neural networks (CNNs) for the identification of ocular diseases using retinal images. We present a comprehensive study on the performance of three optimized CNN architectures Xception, EfficientNet-B3, and ResNet-50 - along with data preprocessing techniques, including augmentation and normalization, to enhance classification accuracy. Experimental results demonstrate that the optimized Xception model achieves the highest classification accuracy of 99%, proving the efficacy of deep learning models in reliable ocular disease detection. The findings emphasize the significance of hyperparameter tuning, optimizer selection, and batch size variations in improving CNN performance for automated ophthalmic diagnosis.