A Hybrid Framework Bridging CNN and ViT based on Theory of Evidence for Diabetic Retinopathy Grading

A new hybrid framework combining convolutional neural networks (CNN) and vision transformers (ViT) has been proposed to enhance the grading of diabetic retinopathy (DR). This innovative approach aims to improve the efficiency of clinical screenings and facilitate early detection of DR, a condition that significantly affects the vision and mental health of many middle-aged and elderly individuals. By addressing the limitations of existing CNN and ViT systems, this research could lead to better outcomes for patients and potentially reduce the burden of vision loss associated with DR.