Retinal Disease Detection and Diagnosis Using Hybrid ResNet-RNN Algorithm

In this study, we propose an improved learning strategy utilising a Hybrid ResNet-RNN model for the early detection of diabetic retinopathy using ResNet-50 for feature extraction and a Recurrent Neural Network (RNN) for classification, enabling an efficient and accurate diagnosis of retinal diseases. To enhance model robustness and practicality, we apply dataset normalization, augmentation, and other preprocessing techniques. The proposed architecture significantly improves precision and responsiveness by integrating transfer learning with sequential pattern recognition. Our empirical results demonstrate a 92% response rate and a 93% accuracy rate, highlighting the potential of AI-driven solutions to revolutionize diagnostic procedures. This study underscores the critical role of early detection in improving patient outcomes and advancing medical imaging technologies.