An Efficient Techniques used for Diagnosis of Mitral Valve Regurgitation Severity
A. Anbarasi1, S. Ravi2, J. Vaishnavi3, S. V. Suresh Babu Matla4
1A. Anbarasi, a Research Scholar in the Department of Computer Science, School of Engineering and Technology, Pondicherry University, Pondicherry, India.
2Dr. S. Ravi, Associate Professor, in the department of Computer Science, School of Engineering and Technology, Pondicherry University, Pondicherry, India.
3J.Vaishnavi, Research Scholor, Department of Computer Science, School of Engineering and Technology, Pondicherry University, Pondicherry, India.
4S. V. Suresh Babu Matla, Research Scholor, Department of Computer Science, School of Engineering and Technology, Pondicherry University, Pondicherry, India.
Manuscript received on October 12, 2019. | Revised Manuscript received on 22 October, 2019. | Manuscript published on November 10, 2019. | PP: 486-491 | Volume-9 Issue-1, November 2019. | Retrieval Number: A4299119119/2019©BEIESP | DOI: 10.35940/ijitee.A4299.119119
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Mitral valve diseases are more common nowadays and might not show up any symptoms. The earlier diagnosis of mitral valve abnormalities such as mitral valve stenosis, mitral valve prolapses and mitral valve regurgitation is most important in order to avoid complex situation. Many existing methodologies such as heart sound investigation model, 3-layered artificial neural network (ANN) of phonocardiogram recordings, 3-layer artificial neural network (ANN) of phonocardiogram recording, echocardiography techniques and so on are there for Mitral Valve diagnosis, but still most of the methods suffer from inefficient image segmentation and misclassification problems. In order to address this issue, this paper proposes two techniques namely 1) Deep Learning based Convolutional Neural Network (CNN) model for Mitral Valve classification model meant for diagnosis and edge detection-based segmentation model to enhance the classifier accuracy. 2) Watershed Segmentation for Mitral Valve identification and image segmentation and Xception model with Random Forest (RF) classifier for training and classification. The proposed models are evaluated in terms of three parameters namely accuracy, sensitivity and specificity, which proved that the proposed models are efficient and appropriate for Mitral Valve diagnosis.
Keywords: Deep Learning Bbased Convolutional Neural Network (CNN) model, Edge Detection-based Segmentation Model, Watershed Segmentation, Xception Model with Random Forest (RF) Classifier.
Scope of the Article: Deep Learning