Infinite Sparse Factor Analysis for Blind Source Separation in Reverberant Environments

  • Kohei Nagira
  • Takuma Otsuka
  • Hiroshi G. Okuno
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7626)

Abstract

Sound source separation in a real-world indoor environment is an ill-formed problem because sound source mixing is affected by the number of sounds, sound source activities, and reverberation. In addition, blind source separation (BSS) suffers from a permutation ambiguity in a frequency domain processing. Conventional methods have two problems: (1) impractical assumptions that the number of sound sources is given, and (2) permutation resolution as a post processing. This paper presents a non-parametric Bayesian BBS called permutation-free infinite sparse factor analysis (PF-ISFA) that solves the two problems simultaneously. Experimental results show that PF-ISFA outperforms conventional complex ISFA in all measures of BSS_EVAL criteria. In particular, PF-ISFA improves Signal-to-Interference Ratio by 14.45 dB and 5.46 dB under RT60 = 30 ms and RT60 = 460 ms conditions, respectively.

Keywords

Blind source separation Reverberant mixtures Infinite sparse factor analysis Non-parametric Bayes 

References

  1. 1.
    Nakadai, K., et al.: Design and Implementation of Robot Audition System ”HARK” Open Source Software for Listening to Three Simultaneous Speakers. Advanced Robotics 24(5-6), 739–761 (2010)CrossRefGoogle Scholar
  2. 2.
    Wölfel, M., et al.: Distant Speech Recognition. Wiley (2009)Google Scholar
  3. 3.
    Sawada, H., et al.: A robust and precise method for solving the permutation problem of frequency-domain blind source separation. IEEE Trans. on Speech and Audio Processing 12(5), 530–538 (2004)CrossRefGoogle Scholar
  4. 4.
    Hyvärinen, A., et al.: Independent component analysis. Wiley Interscience (2001)Google Scholar
  5. 5.
    Sawada, H., et al.: Polar coordinate based nonlinear function for frequency-domain blind source separation. In: Proc. of IEEE Intl. Conf. on Acoustics, Speech, and Signal Processing (ICASSP 2002), pp. 1001–1004 (2002)Google Scholar
  6. 6.
    Lee, I., et al.: Fast fixed-point independent vector analysis algorithms for convolutive blind source separation. Signal Processing 87(8), 1859–1871 (2007)MATHCrossRefGoogle Scholar
  7. 7.
    Hiroe, A.: Solution of Permutation Problem in Frequency Domain ICA, Using Multivariate Probability Density Functions. In: Rosca, J.P., Erdogmus, D., Príncipe, J.C., Haykin, S. (eds.) ICA 2006. LNCS, vol. 3889, pp. 601–608. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  8. 8.
    Nagira, K., Takahashi, T., Ogata, T., Okuno, H.G.: Complex Extension of Infinite Sparse Factor Analysis for Blind Speech Separation. In: Theis, F., Cichocki, A., Yeredor, A., Zibulevsky, M. (eds.) LVA/ICA 2012. LNCS, vol. 7191, pp. 388–396. Springer, Heidelberg (2012)CrossRefGoogle Scholar
  9. 9.
    Sawada, H., et al.: Measuring dependence of bin-wise separated signals for permutation alignment in frequency-domain BSS. In: IEEE Intl. Symposium on Circuits and Systems, ISCAS 2007, pp. 3247–3250. IEEE (2007)Google Scholar
  10. 10.
    Murata, N., et al.: An approach to blind source separation based on temporal structure of speech signals. Neurocomputing 41(1-4), 1–24 (2001)MATHCrossRefGoogle Scholar
  11. 11.
    Griffiths, T., et al.: Infinite latent feature models and the Indian buffet process. Advances in Neural Information Processing Systems 18, 475–482 (2006)Google Scholar
  12. 12.
    Meeds, E., et al.: Modeling dyadic data with binary latent factors. Advances in Neural Information Processing Systems 19, 977–984 (2007)Google Scholar
  13. 13.
    Knowles, D., Ghahramani, Z.: Infinite Sparse Factor Analysis and Infinite Independent Components Analysis. In: Davies, M.E., James, C.J., Abdallah, S.A., Plumbley, M.D. (eds.) ICA 2007. LNCS, vol. 4666, pp. 381–388. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  14. 14.
    Vincent, E., Sawada, H., Bofill, P., Makino, S., Rosca, J.P.: First Stereo Audio Source Separation Evaluation Campaign: Data, Algorithms and Results. In: Davies, M.E., James, C.J., Abdallah, S.A., Plumbley, M.D. (eds.) ICA 2007. LNCS, vol. 4666, pp. 552–559. Springer, Heidelberg (2007)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Kohei Nagira
    • 1
  • Takuma Otsuka
    • 1
  • Hiroshi G. Okuno
    • 1
  1. 1.Graduate School of InformaticsKyoto UniversityKyotoJapan

Personalised recommendations