Optimizing Diabetic Retinopathy Detection with Advanced Machine Learning and Deep Learning Techniques

Diabetes mellitus (DM) is one of the leading causes of severe complications among diabetes patients, diabeteic retinopathy (DR) accounted for 80% up to one in four people with diabetes for more than 15–20 years. The earlier the detection and the sooner the intervention, the less chance there is of vision loss. In this study, machine learning (ML) and deep learning (DL) approaches are demonstrated for DR detection and classification, and their usefulness for advancing healthcare technologies is discussed. Results of the experiments show diagnostic accuracies over 97 % using a CNN-SVD hybrid model with U-net on the EyePACS-1 dataset, and from 98.7% using a DL model based on contrast limited adaptive histogram equalization and generative adversarial networks. However, in some models sensitivity remains a challenge. The findings also highlight the capability of intelligent technologies to accelerate DR diagnosis and away from ophthalmologists, marking the way for integrated public health solutions to prevent diabetes-related blindness. To serve as a resource for researchers exploring innovatve DR detection methods and clinical integration, this review is proposed.