ECG arrhythmia classification using time frequency distribution techniques
In this paper, we focus on classifying cardiac arrhythmias. The MIT-BIH database is used with 14 original classes of labeling which is then mapped into 5 more general classes, using the Association for the Advancement of Medical Instrumentation standard. Three types of features were selected with a focus on the time–frequency aspects of ECG signal. After using the Wigner–Ville distribution the time–frequency plane is split into 9 windows considering the frequency bandwidth and time duration of ECG segments and peaks. The summation over these windows are employed as pseudo-energy features in classification. The “subject-oriented” scheme is used in classification, meaning the train and test sets include samples from different subjects. The subject-oriented method avoids the possible overfitting issues and guaranties the authenticity of the classification. The overall sensitivity and positive predictivity of classification is 99.67 and 98.92%, respectively, which shows a significant improvement over previous studies.
KeywordsCardiac arrhythmia Classification Decision tree Ensemble learner Time–frequency analysis Wigner–Ville distribution
- 1.Jenkins D, Gerred S. Normal ECG. http://www.ecglibrary.com/norm.php. Accessed Febr 2015.
- 3.de Oliveira L, Andreao R, Sarcinelli-Filho M. Premature ventricular beat classification using a dynamic Bayesian network. In: Annual international conference of the IEEE engineering in medicine and biology society, EMBC, Boston, MA; 2011.Google Scholar
- 4.Zeng XD, Chao S, Wong F. Ensemble learning on heartbeat type classification. In: International conference on system science and engineering (ICSSE), Macao; 2011.Google Scholar
- 7.Jiang X, Zhang L, Zhao Q, Albayrak S. ECG arrhythmias recognition system based on independent component analysis feature extraction. In: TENCON 2006 IEEE Region 10 Conference, Hong Kong; 2006.Google Scholar
- 8.Yang S, ShenH. Heartbeat classification using discrete wavelet transform and kernel principal component analysis. In: TENCON Spring Conference, 2013 IEEE, Sydney, NSW; 2013.Google Scholar
- 17.Ghorbani Afkhami R, Tinati MA. ECG based detection of left ventricular hypertrophy using higher order statistics. In: 23rd Iranian Conference on Electrical Engineering (ICEE), Tehran; 2015.Google Scholar
- 20.Zaunseder S, Huhle R, Malberg H. CinC challenge—assessing the usability of ECG by ensemble decision trees. In: Computing in cardiology, Hangzhou; 2011.Google Scholar
- 23.Mark R, Moody G. MIT-BIH database and software catalog. http://ecg.mit.edu/dbinfo.html (1997). Accessed Feb 2015.
- 24.“Recommended practice for testing and reporting performance results of ventricular arrhythmia detection algorithms,” Association for the Advancement of Medical Instrumentation, 1987.Google Scholar
- 25.“Testing and reporting performance results of cardiac rhythm and ST segment measurement algorithms,” Association for the Advancement of Medical Instrumentation, 1998.Google Scholar
- 27.Awodeyi A, Alty S, Ghavami M. Median filter approach for removal of baseline wander in photoplethysmography signals. In: European Modelling Symposium (EMS), Manchester; 2013.Google Scholar
- 28.Prasad G, Sahambi J. Classification of ECG arrhythmias using multi-resolution analysis and neural networks. In: Conference on Convergent Technologies for the Asia-Pacific Region, TENCON, vol. 1, p. 227–231, 2003.Google Scholar
- 29.Martis R, Acharya R, Ray A. Application of higher order cumulants to ECG signals for the cardiac health diagnosis. In: International Conference of the IEEE EMBS, Boston; 2011.Google Scholar
- 30.Ebrahimzadeh A, Khazaee A. Higher order statistics for automated classification of ECG beats. In: International conference on electrical and control engineering (ICECE), Yichang; 2011.Google Scholar