RetinAI-DR Fusion: Dual Deep Learning Framework for Diabetic Retinopathy Detection

Abstract

Diabetic retinopathy (DR) has become one of the leading global causes of vision loss among people living with diabetes. The difficulty lies in its silent progression—tiny changes in the retinal blood vessels often occur long before noticeable symptoms appear, making early detection challenging. Traditional screening approaches rely heavily on expert evaluation of retinal images, which can be time- consuming, costly, and limited in availability, especially in underserved or remote areas. Recent progress in artificial intelligence has begun to change this landscape. Advanced deep learning frameworks such as EfficientNet and ResNet can now analyze retinal fundus images with remarkable speed and precision. These models perform at near-expert accuracy, opening the door to faster, more accessible, and automated screening solutions that can reach larger populations. By increasing efficiency and consistency, AI-driven systems have the potential to improve early diagnosis rates and reduce avoidable blindness. This study explores how these intelligent technologies are revolutionizing diabetic retinopathy detection, emphasizing their potential benefits, current limitations, and the steps needed to successfully integrate them into everyday clinical practice.