4th International Conference on Communication and Electronics Systems (ICCES 2019)

2 / 2019-05-14 14:40:33

Prediction of Cardio Vascular Disease from Retinal Fundus Images Using Neural Networks

Cardio Vascular Disease (CVD), hemorrhages, microaneurysms, exudates, corkscrew arteries, Convolutional Neural Networks (CNN).

Cardio Vascular Disease (CVD) has become the largest single cause of death among humans. Identification and stratification of risk factors of CVD helps in the early detection and treatment of CVD. Retinal fundus images play a significant role in the risk stratification of CVD. The morphological changes in the retinal fundus like hemorrhages, microaneurysms, exudates, and corkscrew arteries are a few of the risk factors. Traditionally, medical discoveries are made by visual exploration, making hypotheses and then designing and running experiments to test the hypotheses. It became extremely tedious because of the wide variety of features, color, patterns, and shapes present in the medical images. In this project, a deep learning model for the prediction of CVD from retinal fundus images has been proposed. The proposed model is trained with the anatomical features of the human eye extracted from the retinal fundus images using image processing techniques. Convolutional Neural Networks (CNN) is used to generate the prediction of CVD. Out of 60 images used, around 53 images were predicted accurately