OCT Vision: Retinal Disease Detection Using Deep Learning

Early diagnosis of retinal diseases such as Age-related Macular Degeneration (AMD), Diabetic Macular Edema (DME), and Retinal Vein Occlusion (RVO) is critical to preserving vision and enabling timely treatment. In this study, we developed a deep learning-based classification system using Optical Coherence Tomography (OCT) images from the OCTDL dataset. We trained the models using TensorFlow, leveraging VGG16 and ResNet50 architectures for comparative performance analysis. VGG16 achieved a classification accuracy of 97.9%, with a recall of 97.5% for AMD, demonstrating its efficacy in distinguishing between retinal conditions. The system is further integrated into a Streamlit-based user interface, allowing real-time predictions with confidence scores, making it accessible for clinical use. Our research highlights the potential of artificial intelligence in enhancing diagnostic accuracy and aiding ophthalmologists in managing retinal diseases.