Back close

Improving the Performance of Cardiac Abnormality Detection from PCG Signal

Publication Type : Conference Paper

Publisher : AIP Conference Proceedings, American Institute of Physics Inc.

Source : AIP Conference Proceedings, American Institute of Physics Inc., Volume 1715 (2016)

Url : https://www.scopus.com/inward/record.uri?eid=2-s2.0-84984539116&partnerID=40&md5=489f68eb9893c87f65627a416925beb7

ISBN : 9780735413627

Campus : Coimbatore

School : School of Engineering

Department : Electronics and Communication

Year : 2016

Abstract : The Phonocardiogram (PCG) signal contains important information about the condition of heart. Using PCG signal analysis prior recognition of coronary illness can be done. In this work, we developed a biomedical system for the detection of abnormality in heart and methods to enhance the performance of the system using SMOTE and AdaBoost technique have been presented. Time and frequency domain features extracted from the PCG signal is input to the system. The back-end classifier to the system developed is Decision Tree using CART (Classification and Regression Tree), with an overall classification accuracy of 78.33% and sensitivity (alarm accuracy) of 40%. Here sensitivity implies the precision obtained from classifying the abnormal heart sound, which is an essential parameter for a system. We further improve the performance of baseline system using SMOTE and AdaBoost algorithm. The proposed approach outperforms the baseline system by an absolute improvement in overall accuracy of 5% and sensitivity of 44.92%

Cite this Research Publication : N. R. Sujit, Dr. Santhosh Kumar C., and Rajesh C. B., “Improving the Performance of Cardiac Abnormality Detection from PCG Signal”, in AIP Conference Proceedings, 2016, vol. 1715.

Admissions Apply Now