Grading and Classification of digital fundus images for Automatic Diabetic retinopathy detection : A Review

Abstract

One of the main causes of blindness and visual impairment in diabetes patients in developed countries isDiabetic Retinopathy (DR), a prominent chronic disease. According to studies, early detection and treatment can help prevent 90% of instances. Physicians employ retinal imaging to test eyes for lesions associated with this condition. The quantity of photos that must be manually evaluated is getting costly due to the rise of diabetic patients. Additionally, it takes a while to train new employees in this form of image-based diagnosis because it necessitates gaining experience through regular practise. In order to identify non-proliferative diabetic retinopathy, numerous methods for spotting micro aneurysms, haemorrhages, and exudates are addressed in this paper. Techniques for detecting blood vessels are also covered for proliferative diabetic retinopathy diagnosis. The report also elaborates on a discussion of the experiments metrics that the scientists used to find diabetic retinopathy. The technical experts and researchers who wish to make use of the current research in this field will benefit from this work.