{"title":"An ICA Algorithm for Separation of Convolutive Mixture of Speech Signals","authors":"Rajkishore Prasad, Hiroshi Saruwatari, Kiyohiro Shikano","volume":23,"journal":"International Journal of Electrical and Computer Engineering","pagesStart":2641,"pagesEnd":2652,"ISSN":"1307-6892","URL":"https:\/\/publications.waset.org\/pdf\/8106","abstract":"
This paper describes Independent Component Analysis (ICA) based fixed-point algorithm for the blind separation of the convolutive mixture of speech, picked-up by a linear microphone array. The proposed algorithm extracts independent sources by non- Gaussianizing the Time-Frequency Series of Speech (TFSS) in a deflationary way. The degree of non-Gaussianization is measured by negentropy. The relative performances of algorithm under random initialization and Null beamformer (NBF) based initialization are studied. It has been found that an NBF based initial value gives speedy convergence as well as better separation performance<\/p>\r\n","references":"[1] P. Common, \"Independent Component Analysis, A New Concept?,\"\r\nSignal Processing,vol.36, pp.287-314, 1994.\r\n[2] A. Hyvarinen, \"Survey on independent component analysis,- Neural\r\nComputing Surveys, vol.2, pp.94-128, 1999.\r\n[3] Cardoso J.F., J. Delabrouille, Guillaume, \"Independent component\r\nanalysis of the cosmic microwave background,\" Proc. ICA2003, 1111-\r\n1116, Nara, Japan, 2003.\r\n[4] Cherry, E. Collin, \"Some experiments on recognition of speech, with\r\none and with two ears,\" Journal of Acoustical Society of America,\r\n25:975-979, 1953.\r\n[5] Cardoso, J.F., \"On the performance of orthogonal source separation\r\nalgorithms,\" Proc. of EUSIPCO-94, Edinburgh, 1994.\r\n[6] Cardoso, J.F., \"Eigenstructure of 4th order cumulants tensor with\r\napplication to the blind source separation problem,\" Proc. ICASSP-89,\r\n2109-2112, 1989.\r\n[7] Jutten, C., Herault, J, \"Blind separation of sources part 1: An adaptive\r\nalgorithm based on neuromimetic architechture,\" J. Signal Processing,\r\n24, 1-10, 1991.\r\n[8] K.Kashino, K. Nakadai, T.Kinoshita, and H.Tanaka, Organization of\r\nhierarchical perceptual sounds,\" Proc. 14th Int. conf. On Artificial\r\nIntelligence, vol.1, 158-164,1995.\r\n[9] T.W. Parson, \"Separation of speech from interfering speech by means of\r\nharmonic selection,\" J. Acoust. Soc. Am., 60, 911-918, 1976.\r\n[10] M. Unoki, M. Akagai, \"A method of signal extraction from noisy signal\r\nbased on auditory scene analysis,\" Speech Communication, 27, 261-279,\r\n1999.\r\n[11] O.L. Frost, \"An algorithm for linearly constrained adaptive array\r\nprocessing,\" Proc. IEEE, 60, 926-935, 1972.\r\n[12] L.J. Griffiths, C.W. Jim, \"An alternative approach to linearly\r\nconstrained adaptive beamforming,\" IEEE Trans. Antennas and\r\nPropag., 30, 27-34, 1982.\r\n[13] Y. Kaneda, J.Ohga, Adaptive microphone array system for noise\r\nreduction,\"IEEE Trans. Acoust., Speech and Signal procs., ASSP-34,\r\n27-34,1986.\r\n[14] H. Saruwatari, S. Kurita, K. Takeda, F. Itakura, T. Nishikawa, and K.\r\nShikano, \"Blind source separation combining independent component\r\nanalysis and beamforming,\" EURASIP Journal on Applied Signal\r\nProcessing,Vol.2003, No.11, pp.1135\u00d4\u00c7\u00f61146, 2003.\r\n[15] T.W.Lee, \"Independent Component Analysis\", Norway, Kluwer\r\nAcademic Press, 1998.\r\n[16] S. Haykin, \"Unsupervised Adaptive Filtering, John Wiley and Sons\r\nInc.,NewYork, 2000.\r\n[17] P. Samaragadis, Blind separation of convolved mixture in the frequency\r\ndomain, Neurocomputing, vol.22, pp.21-34, 1998.\r\n[18] Araki, S. et al., \"The fundamental relation of frequency domain blind\r\nsource separation for convolutive mixures of speech,\". IEEE Trans.\r\nSpeech and Audio Processing, Vol. 11, no.2, 109-116, 2003.\r\n[19] S.Ikeda., S. Murata, \"A method of ICA in time-frequency domain,\"\r\nProc. Workshop Indep. Compon. Anal.Signal Sep., 365-367, 1999.\r\n[20] T.Nishikawa, et al., (2003). Blind Source Separation of Acoustic Signals\r\nBased on Multistage ICA Combining Frequency-Domain ICA and\r\nTime-Domain ICA. IEICE Trans. Fundamentals, Vol.E86-A, pp.846-\r\n58, no.4, April, 2003.\r\n[21] K. Torkkola, \"Blind Separation for audio signals-are we there yet? Proc.\r\nWorkshop on ICA & BSS, France, 1999.\r\n[22] E. Bingham et al., \"A fast fixed point algorithm for independent\r\ncomponent analysis of complex valued signal,\" Int. J. of Neural System,\r\n10(1)1: 8, 2000.\r\n[23] Aapo Hyv\u251c\u00f1rinen, \"Fast and robust fixed-point algorithms for\r\nindependent component analysis,\" IEEE Transactions on Neural\r\nNetworks 10(3):626-634, 1999.\r\n[24] N. Mitianoudis, N. Davies, \"New fixed point solution for convolved\r\naudio source separation,\" Proc. IEEE Workshop on Application of\r\nSignal Processing on Audio and Acoustics, New York. (2001).\r\n[25] Cichocki A., S. Amari, \"Adaptive Blind Signal and Image Processing,\r\nLearning Algorithms and Application,\" John Wiley & Sons Ltd., 130-\r\n131, 2002.\r\n[26] H. Johnson et al, \"Array Signal Processing Concepts and Techniques,\"\r\nPrentice Hall, 1993.\r\n[27] H. Sawada, R. Mukai, S. Araki, S. Makino, A robust approach to the\r\npermutation problem of frequency-domain blind source separation,\"\r\nIEEE International Conference on Acoustics, Speech, and Signal\r\n(ICASSP2003), 381-384, 2003.\r\n[28] S.Kurita,H.Saruwatari,S.Kajita,K.Takeda, and F. Itakura, \"Evaluation\r\nof blind signal separation method using directivity pattern under\r\nreverberant condition,\" Proc. ICASSP2000, vol.5, pp.3140-3143,\r\n(2000).\r\n[29] Godsill S.J., P.J.W. Rayner., \"Digital Audio Restoration,\" Springer\r\nVerlag London, (1998).\r\n[30] Prasad. R. K, H. Saruwatari, K. Shikano, \"Problems in blind separation\r\nof convolutive speech mixture by negenotropy maximization,\", Proc.\r\nIWAENC 2003, Kyoto, Japan, 287-290. (2003)","publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 23, 2008"}