Feature Extraction Based on Pan Tompkins Algorithm from ECG Signals and Diagnosis of Arrhythmia Using Multilayer Perceptron Neural Network

Year 2016, Volume: 5 , 205 - 219, 07.11.2016

Ersin Ersoy Mahmut Hekim

Abstract

In this study, Electrocardiogram (ECG) signals giving information about the state and functioning of the heart are divided into segments, waves and intervals by resting upon temporal limitations and feature vector of each section is obtained by means of arithmetic mean which is one of basic statistical parameters. Arrhythmia (rhythm disorders) occurring in the heart are diagnosed by the obtained feature vectors used as the inputs into multilayer perceptron neural network (MLPNN) model. For this purpose, ECG signals are divided into sections that are 10-minute-equal-length. These sections are divided into subsections (segment and intervals) which are admitted for each segment and wave interval and give information on arrhythmia by temporal limitations and arithmetic average of each interval is used as the inputs into the model of MLPNN for the diagnosis of arrhythmia. As a conclusion, it is proved that the proposed approach has reached to high accuracy rates of classification for the diagnosis of arrhythmia through ECG signals.

Keywords

ECG signals, multilayer perceptron neural network, signal processing, diagnosis of disease, arrhythmia diagnosis

References

• F. Aydin, Developing a machine learning based system to assist treatment processes of arrhythmia patients, Trakya University Institute of Science and Computer Engineering, 2011.
• B. Gumus, S. Yazgi, Automatic detection of cardiac arrhythmia ECG signals using artificial neural network, In: Proceeding of Electrical-Electronics-Computer and Biomedical Engineering 13th National Congress, 23-26 2009 (in Turkish).
• U. Aksu, Comparison of standard and lewis ECG in detection of A-V dissociation in patients with wide QRS tachycardia, Atatürk University Faculty of Medicine, Department of Cardiology, 2014.
• N. Ozcan, A fuzzy support vector machine approach for ECG analysis, Bogazici University in Systems and Control Engineering, 2010.
• M. Korueki, S. Metin, Classification of electrocardiogram beats using GAL network, Istanbul Technical University, Institute of Science and Technology, 2002.
• Z. Dokur, Classification of ECG beats by using artificial neural networks and genetic algorithms, Istanbul Technical University Institute of Science Electrical and Electronics Engineering Department, 1999.
• B. Ilerigelen, H. Mutlu, Continuing Medical Education Faculty of Cerrahpaşa Medical Techniques, ECG course booklet, http://www.ctf.edu.tr/stek/EKG_Kurs_Kitap.pdf (in Turkish).
• Physionet, MIT-BIH Normal sinus rhythm Database http://www.physionet.org/physiobank/database/nsrdb/ (May 10, 2016).
• Physionet, MIT-BIH Arrhythmia Database, http://www.physionet.org/physiobank/database/mitdb/ (May 10, 2016).
• S.L. Pingale, N. Daimiwal, Detection of various diseases using ECG signal in Matlab, International Journal of Recent Technology and Engineering. ISSN: 2277-3878, 2014.
• J.H. Indik, E.C. Pearson, K. Fried, R.L. Woosley, B. Fridericia, QT correction formulas interfere with measurement of drug-induced changes in QT interval, Sarver Heart Center of University of Arizona College of Medicine, 2006.
• J. Pan, W.J. Tompkins, A real-time QRS detection algorithm, IEEE Transactions on Biomedical Engineering, BME-32 (3) 230-236, 1985.
• C.M. Bishop, Neural Networks for Pattern Recognition, Oxford University Press Inc. New York, NY, USA 1995.
• M.W. Gardner, S.R. Dorling, Artificial neural networks (the multilayer perceptron)—a review of applications in the atmospheric sciences, School of Environmental Sciences, University of East Anglia, Norwich, Norfolk NR4 7TJ, 1998.
• M. Hekim, U. Orhan, M. Ozer, Epileptic seizure detection using artificial neural network and a new feature extraction approach based on equal width discretization, Gazi University Journal of the Faculty of Engineering and Architecture, 26(3) 575-580, 2011.