Feature Extraction Based on Optimal Discrimination Plane in ECG Signal Classification

  • Dingfei Ge
  • Xiao Qu
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4093)


In order to improve the classification results on electrocardiogram (ECG) signals, Optimal Discrimination Plane (ODP) approach is introduced. Features are extracted from time-series data using the ODP that is developed by Fisher’s criterion method. ECG patterns are projected onto two orthogonal vectors, and the two-dimensional feature vectors are used as features to represent the ECG segments. Two types of ECG signals are obtained from MIT-BIH database, namely normal sinus rhythm and premature ventricular contraction. A quadratic discriminant function based classifier and a threshold vector based classifier are employed to classify these ECG beats, respectively. The results show the proposed technique can achieve better classification results compared to that of some recently published on arrhythmia classification.


Principle Component Analysis Normal Sinus Rhythm Premature Ventricular Contraction Quadratic Discriminant Function Threshold Vector 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Khadra, L., al-Fahoum, A.S., al-Nashash, H.: Detection of life-threatening cardiac arrhythmias using the wavelet transformation. Med. Biol. Eng. Comput. 35, 626–632 (1997)CrossRefGoogle Scholar
  2. 2.
    Mroczka, T., Lewandowski, P., Maniewski, R., et al.: Effectiveness of high resolution ECG spectral analysis in discrimination of patients prone to ventricular tachycardia and fibrillation. Med. Sci. Monit. 6, 1018–1026 (2000)Google Scholar
  3. 3.
    Mohamed, I.O., Abou-Zied Ahmed, H., Youssef Abou-Bakr, M., et al.: Study of features based on nonlinear dynamical modeling in ECG arrhythmia detection and classification. IEEE Trans. Biomed. Eng. 49, 733–736 (2002)CrossRefGoogle Scholar
  4. 4.
    Ge, D.F., Srinivasan, N., Krishnan, S.M.: Cardiac arrhythmia classification using autoregressive modeling. Biomedical Engineering Online 1, 5 (2002)CrossRefGoogle Scholar
  5. 5.
    Sun, Y., Chan, K.L., Krishnan, S.M.: Life-threatening ventricular arrhythmia recognition by nonlinear descriptor. Biomed. Eng. Online 4, 6 (2005)CrossRefGoogle Scholar
  6. 6.
    Christon, I.I.: Real time electrocardiogram QRS detection using combined adaptive threshold. Biomed. Eng. Online 1, 28 (2004)CrossRefGoogle Scholar
  7. 7.
    Zhou, S.H., Rautaharju, P.M., Calhoun, H.P.: Selection of a reduced set of parameters for classification of ventricular conduction defects by cluster analysis. Proc. Comp. Cardiol., 879–882 (1993)Google Scholar
  8. 8.
    Ge, D.F., Shao, Y.Q., Jiang, H.Z.: An algorithm study on telecardiogram diagnosis based on multivariate autoregressive model and two-lead ECG signals. Space Med. Med. Eng. 17, 355–359 (2004)Google Scholar
  9. 9.
    Fukunaga, K.: Introduction to statistical pattern recognition. Academic Press Limited, United States of America, New York (1990)MATHGoogle Scholar
  10. 10.
    Duda, R.O., Hart, P.E.: Pattern classification. Wiley Interscience Publication, New York (2001)MATHGoogle Scholar
  11. 11.
    Melo, S.L., Caloba, L.P., Nadal, J.: Arrhythmia analysis using artificial neural network and decimated electrocardiographic data. Comp. Cardiol. 27, 73–76 (2000)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Dingfei Ge
    • 1
  • Xiao Qu
    • 1
  1. 1.School of Information and Electronic EngineeringZhejiang University of Science and TechnologyHangzhouP.R.C

Personalised recommendations