Advanced Deep Learning Techniques for Comprehensive Detection of Eye Disease Using Retinal and OCT Imaging

Early detection of eye diseases are crucial for preventing vision loss and ensuring timely treatment. This paper explores the application of advanced deep learning techniques for the comprehensive detection of various eye diseases using retinal and Optical Coherence Tomography (OCT) imaging. The detection of eye diseases, particularly myopia, is an important healthcare challenge in Malaysia due to the increasing prevalence of vision-related disorders. This research focuses on developing an AI-driven solution to address this challenge, with the primary focus on detecting myopia. However, the system is also capable of identifying other conditions such as acrima, retinal diseases, origa, diabetic retinopathy, cataract, glaucoma and age-related macular degeneration. The study utilizes Convolutional Neural Networks, achieving a high accuracy of 97.87% for myopia detection. Fine-tuning was applied to a pre-trained CNN model, leveraging transfer learning to enhance the model's performance. By employing advanced deep learning architectures, this research enhances diagnostic accuracy and efficiency, providing a robust framework for the detection of a wide range of ocular diseases. The results highlight the potential of CNNs in revolutionizing eye care, emphasizing the role of AI in improving diagnostic capabilities and its integration with retinal and OCT imaging to ensure timely diagnosis and treatment.