DIABETIC RETINOPATHY ANALYSIS AND DETECTION USING DEEP LEARNING

Abstract

Diabetic Retinopathy is a disease that needs to be detected early because without doctor intervention it can progress to blindness. The traditional method of examination is very time-consuming and not reachable by everyone who suffers from diabetes. Since the number of patients is increasing, researchers have been studying various techniques to improve the detection of the disease even when using non- professional cameras. Artificial Intelligence has had great improvements and it is becoming widely used in medicine as well. Various Deep Learning techniques have been used and the results achieved are admirable, yet some of them are not feasible to be applied in real life. The purpose of this study is to compare different Machine Learning and Deep Learning techniques used for the detection and analysis of Diabetic Retinopathy, as well as optimize a model not only in terms of results, but in terms of the generalization of the model and the computational power it uses by using hyperparameter optimization techniques and comparing different optimizers.