Artificial Intelligence and Machine Learning Applications in Diabetic Retinopathy Screening and Glaucoma Detection: A Review Article

Artificial intelligence (AI) and machine learning (ML) have become groundbreaking technologies in the field of Ophthalmology and especially in the screening of diabetic retinopathy (DR) and glaucoma. This review discusses how AI/ML systems are currently applied, methodologies used, performance metrics, and clinical implementation issues of AI/ML systems in the detection of these vision-threatening conditions. Very recent deep learning algorithms have achieved the diagnostic accuracy of human experts, or more, with a sensitivity and specificity rate over 90% in many of the studies. However, there are challenges such as dataset bias, regulatory approval, clinical integration and cost-effectiveness that need to be investigated further. This review is a synthesis of evidence obtained from recent literature and discusses future directions of AI powered ophthalmic diagnostics.