Implementation of Vision Transformers (ViTs) based Advanced Iris Image Analysis for Non-Invasive Detection of Diabetic Conditions

Diabetes mellitus is a prevalent metabolic disorder that requires early detection to prevent severe complications. Traditional diagnostic methods, such as blood tests, are often invasive, costly, and impractical for routine monitoring. This study proposes a novel, noninvasive diagnostic approach leveraging advanced iris image analysis with Vision Transformers (ViTs). By capturing subtle structural and textural variations in the iris associated with diabetes, our method enables accurate and efficient disease identification. The proposed framework involves preprocessing techniques to enhance crucial iris features, followed by feature extraction and classification using a fine-tuned ViT model. ViTs excel in capturing long-range dependencies and complex patterns, making them well-suited for detecting diabetic indicators within iris textures. Our model is trained and validated on a diverse dataset comprising iris images from both diabetic and non-diabetic individuals, ensuring robustness and generalizability. This non-invasive approach, combined with the interpretability and scalability of ViTs, presents a transformative alternative to conventional diagnostic techniques. By integrating AI-driven healthcare solutions, this research contributes to the advancement of early diabetic detection, offering a practical and accessible tool for proactive disease management.