Publication Type : Conference Paper
Publisher : 2019 3rd International Conference on Trends in Electronics and Informatics (ICOEI)
Source : 2019 3rd International Conference on Trends in Electronics and Informatics (ICOEI), IEEE, Tirunelveli, India, India (2019)
Url : https://ieeexplore.ieee.org/document/8862719
ISBN : 9781538694398
Keywords : Arteries, Artery Vein Ratio, Automated diagnosis, AV classification, biomedical optical imaging, Blood vessels, common condition, common retinal diseases, Convolution neural network, Diseases, Eye, Feature extraction, Feature vectors, heart rate, hypertensive retinopathy, image classification, Image segmentation, Medical Image Processing, neural network, Radon vessel tracking algorithm, Retina, retinal disorders, Retinal vasculature, Retinal Vessels, Support Vector Machine, Support vector machines, sustained hypertension, Veins
Campus : Coimbatore
School : School of Computing, School of Engineering
Department : Computer Science
Year : 2019
Abstract : Hypertensive Retinopathy (HR) is one of the most common retinal diseases. Hypertension diversely affects different body parts, including the eyes. Sustained hypertension can lead to damage in the retinal vasculature causing vision problems, this condition is termed as Hypertensive Retinopathy (HR). Since HR is a common condition associated with several Cardio-Vascular Diseases (CVD), automated diagnosis of HR can aid the physician when dealing with a large population. Classification of retinal vessels is the first step in characterizing retinal disorders such as HR. In this work, we have proposed an automated support system for classification of arteries and veins (AV) for HR detection. The proposed framework classifies AV using different feature vectors obtained through Radon vessel tracking algorithm. The features are extracted from publicly available standard dataset DRIVE. One of the main advantages of vessel tracking is that it can be further utilized for detecting Artery Vein Ratio (AVR). Classification results are acquired using three different classifiers namely SVM, NN and CNN. Experimental results show that CNN outperforms NN and SVM in AV classification for HR detection.
Cite this Research Publication : M. Kiruthika, Swapna, T. R., Santhosh, K. C., and Peeyush, K. P., “Artery and Vein classification for hypertensive retinopathy”, in 2019 3rd International Conference on Trends in Electronics and Informatics (ICOEI), Tirunelveli, India, India, 2019.