OPTIMIZING EYE DISEASE DIAGNOSIS ALGORITHMS THROUGH PARALLEL PROCESSING

The accurate diagnosis of eye diseases is crucial for preventing severe visual impairments. In this research, I personally propose the optimization of eye disease classification algorithms through the use of parallel processing. More precisely, I utilize deep learning techniques, specifically the EfficientNetV2-B0 model, for the classification of fundus images into four categories: cataract, diabetic retinopathy, glaucoma, and normal. I compare the performance of a sequential model against a parallel multi-GPU model to evaluate the benefits of parallel processing. Research results show a significant reduction in training time with minimal impact on classification accuracy, demonstrating the potential of parallelism in optimizing the computational efficiency of eye disease diagnosis systems.