Enhancing Retinal Disease Classification with Dual Scale Twin Vision Transformers using OCT Imaging

Retinal optical coherence tomography (OCT) imaging has become a valuable tool in diagnosing and monitoring retinal diseases. Recent advancements in vision transformers have demonstrated impressive performance in image classification tasks. This research article presents a novel approach, named Dual Scale Twin Vision Transformer (DST-ViT), for retinal disease classification using OCT images.DST-ViT combines the advantages of dual-scale representation learning and the Twin transformer architecture to improve disease classification accuracy. By leveraging dual-scale input OCT images, DST-ViT effectively captures fine-grained details and global contextual information simultaneously, enhancing classification performance. With an outstanding average accuracy of 99.92%, DST-ViT surpasses existing state-of-the-art methods.This research contributes to the growing understanding of ViT and their application in retinal disease classification. The promising results achieved by DST-ViT open possibilities for improved diagnosis and treatment planning in the field of ophthalmology.