Automated neural network detection of wavelet preprocessed electrocardiogram late potentials

  • A. Rakotomamonjy
  • B. Migeon
  • P. Marche


The aim of the study is to investigate the potential of a feedforward neural network for detecting wavelet preprocessed late potentials. The terminal parts of a simulated QRS complex are processed with a continuous wavelet transform, which leads to a time-frequency represenation of the QRS complex. Then, diagnostic feature vectors are obtained by subdividing the representations into several regions and by processing the sum of the decomposition coefficients belonging to each region. The neural network is trained with these feature vectors. Simulated ECGs with varying signalto-noise ratios are used to train and test the classifier. Results show that correct classification ranges from 79% (high-level noise) to 99% (no noise). The study shows the potential of neural networks for the classification of late potentials that have been preprocessed by a wavelet transform. However, clinical use of this method still requires further investigation.


Late potentials Neural networks Wavelet transform 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Atarius, R., andSornmo, L. (1995a): ‘Signal to noise enhancement of cardiac late potential using ensemble correlation’,IEEE Trans.,BME-42, pp. 1132–1137Google Scholar
  2. Atarius, R. andSornmo, L. (1995b): ‘Detection and estimation of cardiac late potentials in non-stationary noise.’ Dept. Elec. Eng., Lund., Tech. Report SPR 27Google Scholar
  3. Berbari, E. J. (1987): ‘Critical overivew of late potentials recording’,J. Electrocardiogr., Oct., pp. 125–127Google Scholar
  4. Bortolan, G., Degani, R. andWillems, J. L. (1990): Design of neural networks for classification of electrocardiography signals.’ Proc. 12th Annual Int. Conf. IEEE-EMBS, Philadelphia, USAGoogle Scholar
  5. Doncarli, C., Goerig, L. andAuger, F. (1990): ‘Detection of late potentials in ECG by means of an adaptive smoother wavelet transfor.’ Proc. EUSIPOGoogle Scholar
  6. Fuller, M. S., Dustman, T. andFreedman, R. (1991): ‘Wavelet analysis of signal averaged electrocardiogram,’Electrocardiography, pp. 639–640Google Scholar
  7. Grossman, A., andMorlet, J. (1984): ‘Decomposition of hardy functions into square integrable wavelets of constant shape,’SIAM J. Math. Anal.,15, (4), pp. 723–736MathSciNetCrossRefGoogle Scholar
  8. Guo, Z., Durand, L., Lee, H., Allard, L., Grenier, M. andStein, P. (1994): ‘Artificial neural networks in computer-assisted classification of heart sounds in patients with porcine bioprosthetic valves’,Med. Biol. Eng. Comput.,32, (5), pp. 311–316CrossRefGoogle Scholar
  9. Hush, D. andHorne, B. (1993): ‘Progress in supervised neural networks: What’s new since Lippman?’IEEE SP Mag.,1, pp. 8–39Google Scholar
  10. Kelly, K., Escalona, O. J. andMitchell, R. H. (1992): ‘The use of adaptive line enhancement for beat to beat detection of late potentials: an evaluation’,Innov. Tech. Biol. Med.,13, pp. 587–600Google Scholar
  11. Kronland-Martinet, R., Morlet, J. andGrossman, A. (1987): ‘Analysis of sound patterns through wavelet transforms’,Int. J. Pattern. Recognit. Art. Intell.,1, (2), pp. 273–302CrossRefGoogle Scholar
  12. Lander, P., andBerbari, E. J. (1989): ‘Use of high-pass filtering to detect late potentials in the signal-averaged ECG’,J. Electrocardiol,22, (suppl.), pp. 7–12Google Scholar
  13. Lippmann, R. (1987): ‘An introduction to computing with neural nets’,IEEE ASSP Mag., April, pp. 7–22Google Scholar
  14. Meste, O. andRix, H. (1989): ‘Detection of late potentials by means of wavelet transform’. Proc. 11th Annual Int. Conf. IEEE-EMBS, Seattle, USAGoogle Scholar
  15. Meste, O., Rix, H., Carminal, P., andThakor, V. N. (1994): ‘Ventricular late potentials characterization in time-frequency domain by means of a wavelet transform’,IEEE Trans.,BME-41, pp. 625–633Google Scholar
  16. Rioul, O., andVetterli, M. (1991): ‘Wavelets and signal processing’,IEEE Signal Process. Mag.,8, pp. 14–35CrossRefGoogle Scholar
  17. Sabbatini, R. M. E. (1993): ‘Neural networks for classification and pattern recognition of biological signals.’ Proc. 15th Annual Int. Conf. IEEE-EMBS, pp. 265–266Google Scholar
  18. Sehadjit, L., Belanger, J. J., Carrault, G. andCoatgrieux, J.L. (1990): ‘Wavelet analysis of ECG signal.’ Proc. 12th Annual Int. Conf. IEEE-EMBS Philadelphia, USA, pp. 811–812Google Scholar
  19. Simon, M. B. (1981). ‘Use of signals in the terminal QRS complex to identify patients with ventricular tachycardia after myocardial infarction’,Circulation,64, pp. 234–242Google Scholar
  20. Sornmo, L. andAtarius, R. (1995): ‘Effects of noise in maximum likelihood analysis of late potentials,’J. Electrocardiol.,28, (suppl.)Google Scholar
  21. Suzuki, Y. andOno, K. (1992): ‘Personal computer system for ECG-ST segment recognition based on neural network’,Med. Biol. Eng. Comput.,30, (1), pp. 2–8CrossRefGoogle Scholar
  22. Svensson, O. andSornmo, L. (1992): ‘Subband analysis of cardiac late potentials’. Proc. 14th Annual Int. Conf. IEEE-EMBS, pp. 492–493Google Scholar
  23. Tuteur, F. B. (1988): ‘Wavelet transformation in signal detection’,IEEE ICASSP,3, pp. 1435–1438Google Scholar
  24. Xue, Q. andReddy, B. R. S. (1993): ‘Late potential recognition by artificial neural networks.’ Proc. 15th Annual Int. Conf. IEEE-EMBS, pp. 717–718Google Scholar

Copyright information

© IFMBE 1998

Authors and Affiliations

  1. 1.Laboratoire Vision et RobotiqueInstitut Universitaire de TechnologieBourgesFrance

Personalised recommendations