×

Efficient blind dereverberation and echo cancellation based on independent component analysis for actual acoustic signals. (English) Zbl 1238.94015

Summary: This letter presents a new algorithm for blind dereverberation and echo cancellation based on independent component analysis (ICA) for actual acoustic signals. We focus on frequency domain ICA (FD-ICA) because its computational cost and speed of learning convergence are sufficiently reasonable for practical applications such as hands-free speech recognition. In applying conventional FD-ICA as a preprocessing of automatic speech recognition in noisy environments, one of the most critical problems is how to cope with reverberations. To extract a clean signal from the reverberant observation, we model the separation process in the short-time Fourier transform domain and apply the multiple input/output inverse-filtering theorem (MINT) to the FD-ICA separation model. A naive implementation of this method is computationally expensive, because its time complexity is the second order of reverberation time. Therefore, the main issue in dereverberation is to reduce the high computational cost of ICA. In this letter, we reduce the computational complexity to the linear order of the reverberation time by using two techniques: (1) a separation model based on the independence of delayed observed signals with MINT and (2) spatial sphering for preprocessing. Experiments show that the computational cost grows in proportion to the linear order of the reverberation time and that our method improves the word correctness of automatic speech recognition by 10 to 20 points in an \(\mathrm{RT}_{20}= 670\) ms reverberant environment.

MSC:

94A12 Signal theory (characterization, reconstruction, filtering, etc.)
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] DOI: 10.1162/089976698300017746 · doi:10.1162/089976698300017746
[2] DOI: 10.1093/ietfec/e88-a.12.3593 · doi:10.1093/ietfec/e88-a.12.3593
[3] DOI: 10.1109/TSA.2003.809193 · doi:10.1109/TSA.2003.809193
[4] DOI: 10.1155/S1110865704402303 · Zbl 02179541 · doi:10.1155/S1110865704402303
[5] DOI: 10.1007/978-0-387-45528-0 · doi:10.1007/978-0-387-45528-0
[6] DOI: 10.1109/TASSP.1979.1163209 · doi:10.1109/TASSP.1979.1163209
[7] Choi S., Proc. of International Workshop on ICA and BBS pp 371– (1999)
[8] DOI: 10.1109/TSA.2004.838538 · doi:10.1109/TSA.2004.838538
[9] DOI: 10.1109/TASL.2007.898456 · doi:10.1109/TASL.2007.898456
[10] DOI: 10.1016/S0165-1684(01)00101-3 · Zbl 1076.94520 · doi:10.1016/S0165-1684(01)00101-3
[11] DOI: 10.1109/89.876299 · doi:10.1109/89.876299
[12] DOI: 10.1109/89.279278 · doi:10.1109/89.279278
[13] DOI: 10.1016/S0165-1684(00)00049-9 · doi:10.1016/S0165-1684(00)00049-9
[14] DOI: 10.1109/ICASSP.2008.4518676 · doi:10.1109/ICASSP.2008.4518676
[15] Haykin S., Adaptive filter theory (1991) · Zbl 0723.93070
[16] DOI: 10.1109/TASL.2007.894530 · doi:10.1109/TASL.2007.894530
[17] DOI: 10.1007/978-3-540-74494-8_59 · Zbl 1173.94364 · doi:10.1007/978-3-540-74494-8_59
[18] DOI: 10.1109/78.790650 · doi:10.1109/78.790650
[19] DOI: 10.1002/0471221317 · doi:10.1002/0471221317
[20] Ikeda S., Proc. Int’l Workshop on Independent Component Analysis and Signal Separation pp 365– (1999)
[21] DOI: 10.1109/97.935740 · doi:10.1109/97.935740
[22] DOI: 10.1109/TSA.2005.854109 · doi:10.1109/TSA.2005.854109
[23] DOI: 10.1109/TSP.2006.872545 · Zbl 1373.94637 · doi:10.1109/TSP.2006.872545
[24] DOI: 10.1006/csla.1995.0010 · doi:10.1006/csla.1995.0010
[25] DOI: 10.1109/89.221372 · doi:10.1109/89.221372
[26] DOI: 10.1155/2007/57470 · Zbl 1168.94428 · doi:10.1155/2007/57470
[27] DOI: 10.1109/29.1509 · doi:10.1109/29.1509
[28] DOI: 10.1016/S0925-2312(00)00345-3 · Zbl 0998.68134 · doi:10.1016/S0925-2312(00)00345-3
[29] DOI: 10.1109/ICASSP.2003.1198724 · doi:10.1109/ICASSP.2003.1198724
[30] Nishikawa T., IEICE Trans. Fundamentals 86 pp 846– (2003)
[31] DOI: 10.1109/MSP.2005.1511828 · doi:10.1109/MSP.2005.1511828
[32] Rosenthal D., Computational auditory scene analysis (1998)
[33] DOI: 10.1109/TSA.2005.855832 · doi:10.1109/TSA.2005.855832
[34] Sawada H., IEICE Trans. Fundamentals 86 pp 505– (2003)
[35] DOI: 10.1109/TAP.1986.1143830 · doi:10.1109/TAP.1986.1143830
[36] DOI: 10.1109/97.736233 · doi:10.1109/97.736233
[37] Vaidyanathan P. P., Multirate systems and filter banks (1993) · Zbl 0784.93096
[38] DOI: 10.1109/TCSII.2008.2008060 · doi:10.1109/TCSII.2008.2008060
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.