A Robust Correlation Method for Solving Permutation Problem in Frequency Domain Blind Source Separation of Speech Signals

This paper addresses the well-known permutation problem, inconsistent permutation in the discrete Fourier transform bins after independent component analysis, in the frequency domain blind source separation. The method which utilizes the correlation between the adjacent bins in speech signals is popular among the techniques for solving permutation problem. However, the reliability of the method is very low. This paper presents a robust correlation method for solving the permutation problem, which utilizes the correlation between the signals in each DFT bin, one of which is partially separated by a time domain BSS method and the other is obtained by a frequency domain BSS method. The proposed method showed almost the same separation performance as that of the method which uses the correlation between adjacent bins and is highly reliable at the same time

[1]  Te-Won Lee,et al.  Independent Component Analysis , 1998, Springer US.

[2]  Dennis R. Morgan,et al.  A beamforming approach to permutation alignment for multichannel frequency-domain blind speech separation , 2002, 2002 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[3]  Walter Kellermann,et al.  Real-Time Convolutive Blind Source Separation Based on a Broadband Approach , 2004, ICA.

[4]  Nobuhiko Kitawaki,et al.  A combined approach of array processing and independent component analysis for blind separation of acoustic signals , 2001, 2001 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.01CH37221).

[5]  Paris Smaragdis,et al.  Blind separation of convolved mixtures in the frequency domain , 1998, Neurocomputing.

[6]  西川 剛樹.  Blind source separation based on multistage independent component analysis , 2005 .

[7]  Andreas Ziehe,et al.  An approach to blind source separation based on temporal structure of speech signals , 2001, Neurocomputing.

[8]  Aapo Hyvärinen,et al.  A Fast Fixed-Point Algorithm for Independent Component Analysis of Complex Valued Signals , 2000, Int. J. Neural Syst..

[9]  Hiroshi Sawada,et al.  A robust and precise method for solving the permutation problem of frequency-domain blind source separation , 2004, IEEE Transactions on Speech and Audio Processing.

[10]  Kazuya Takeda,et al.  Evaluation of blind signal separation method using directivity pattern under reverberant conditions , 2000, 2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100).