Prediction diabetic retinopathy from retinal fundus images via artificial neural network

Abstract

Diabetic Retinopathy (DR) is a vascular disorder affecting the retina due to prolonged diabetes. DR is one of the causes of vision defection. The risk of blindness can be reduced significantly in patients by early screening of diabetic patients for the development of diabetic retinopathy. In this study, vascular structure was extracted from fundus images using the Kirsch's method. Gaussian filter has been applied to the extracted vascular structure and removed from its noise. In this way, the detection of DR by pre-trained artificial intelligence has been made significantly easier.