Speech Separation Based on Improved Fast ICA with Kurtosis Maximization of Wavelet Packet Coefficients

Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 275)

Abstract

To improve the separation performance of ICA algorithm, wavelet packets transformation was adopted to reduce the signals’ overlapped degree, that was, the mixture speech signals were decomposed into wavelet packets, and the node that had the highest kurtosis was the optimal wavelet packets decomposition node since the kurtosis is a measure of non-Gaussian nature. Thereby, it reduced the signals’ overlapped degree in the wavelet domain. Then the separation matrix was calculated by using FastICA algorithm iteratively, and the source signal estimations were obtained finally. Simulation results demonstrated the separation performance improved clearly when compared with FastICA algorithm in time domain and other wavelet FastICA method.

Keywords

wavelet packet FastICA speech separation optimal wavelet packets decomposition 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Yu, X.C., Hu, D.: Blind source separation theory and application, Beijing (2011)Google Scholar
  2. 2.
    Zhao, C.H., Liu, J., Sun, J.D.: Blind separation of noisy speech mixtures based on wavelet transform and independent component analysis. Joumal of Electronics&Infbrnlation Tcchnology 28(9), 1565–1568 (2006)Google Scholar
  3. 3.
    Zhao, T.J., He, X.S., Chen, L.: Nosiy blind source separation algorithm based on new threshold function of wavelet transform. Application Research of Computers 27(8), 2886–2888 (2010)Google Scholar
  4. 4.
    Liao, T., Liu, H.: Improved step-size adaptive method of wavelet ICA image blind separation. Computer Engineering and Applications 47(32), 221–223 (2011)Google Scholar
  5. 5.
    Zhang, C., Yang, J.A., Ye, F.: Single channel blind separation algorithm based on singular value decomposition of wavelet components. Journal of Electronic Measurement and Instrument 25(11), 991–997 (2011)Google Scholar
  6. 6.
    Dario, F., Marie, F.L., Christian, D.: Optimized wavelets for blind se paration of nonstationary surface myoelectric signals. IEEE Transactions on Biomedical Engineer 55(1), 78–86 (2008)Google Scholar
  7. 7.
    Moussaoui, R., Rouat, J., Lefebvre, R.: Wavelet based independent component analysis for multi-channel source separation. In: IEEE International Conference on Acoustic, Speech and Signal Processing, vol. 5, pp. V645–V648 (2006)Google Scholar
  8. 8.
    Hyvarinen, A.: Fast and robust fixed-point algorithms for independent component analysis. IEEE Transactions on Neural Computation 10(3), 626–634 (1999)CrossRefGoogle Scholar
  9. 9.
    Wang, D.K., Peng, J.Y.: Wavelet analysis and its applications in signal processing, Beijing (2006)Google Scholar
  10. 10.
    Emmanuel, V., Remi, G., Cedric, F.: Performance measurement in blind audio source separation. IEEE Transactions on Audio, Speech, and Language Processing 14(4), 1462–1469 (2006)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  1. 1.Key Laboratory of Advanced Process Control for Light IndustryJiangnan UniversityWuxiChina

Personalised recommendations