Lessons learnt from harnessing deep learning for real-world clinical applications in ophthalmology: detecting diabetic retinopathy from retinal fundus photographs

Abstract Diabetic retinopathy (DR) is one of the fastest growing causes of blindness and has prompted the implementation of national screening programs. To help address the shortage of experts to grade images for signs of DR, there has been a surge of interest in artificial intelligence for DR detection. In this chapter, we will cover both historical and recent deep learning algorithms for automated DR detection, the current state of regulatory approval and clinical validation, and future outlook.