Underdetermined Blind Source Separation of Convolutive Mixtures by Hierarchical Clustering and L1-Norm Minimization

  • Stefan Winter
  • Walter Kellermann
  • Hiroshi Sawada
  • Shoji Makino
Part of the Signals and Communication Technology book series (SCT)

In this chapter we present a complete solution for underdetermined blind source separation (BSS) of convolutive speech mixtures based on two stages. In the first stage, the mixing system is estimated, for which we employ hierarchical clustering. Based on the estimated mixing system, the source signals are estimated in the second stage. The solution for the second stage utilizes the common assumption of independent and identically distributed sources. Modeling the sources by a Laplacian distribution leads to ℓ1-norm minimization.

Keywords

Microwave Assure Acoustics Estima Maki 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    O. Yilmaz and S. Rickard, “Blind separation of speech mixtures via time-frequency masking,” IEEE Transactions on Signal Processing, vol. 52, no. 7, pp. 1830-1847, July 2004. [Online]. Available: http://eleceng.ucd.ie/~srickard/bss.html
  2. 2.
    S. Rickard and O. Yilmaz, “On the approximate W-disjoint orthogonality of speech,” in Proc. ICASSP 2002, vol. 1, 2002, pp. 529-532.Google Scholar
  3. 3.
    L. Vielva, I. Santamaria, C. Pantaleon, J. Ibanez, and D. Erdogmus, “Estima-tion of the mixing matrix for underdetermined blind source separation using spectral estimation techniques,” in Proc. EUSIPCO 2002, vol. 1, Sept. 2002, pp. 557-560.Google Scholar
  4. 4.
    P. Bofill and M. Zibulevsky, “Blind separation of more sources than mixtures using sparsity of their short-time Fourier transform,” in Proc. ICA 2000, June 2000, pp. 87-92.Google Scholar
  5. 5.
    P. Bofill, “Underdetermined blind separation of delayed sound sources in the frequency domain,” Neurocomputing, vol. 55, no. 3-4, pp. 627-641, Oct. 2003.CrossRefGoogle Scholar
  6. 6.
    S. Araki, S. Makino, A. Blin, R. Mukai, and H. Sawada, “Underdetermined blind separation for speech in real environments with sparseness and ICA,” in Proc. ICASSP 2004, vol. III, May 2004, pp. 881-884.Google Scholar
  7. 7.
    A. Blin, S. Araki, and S. Makino, “Underdetermined blind separation of convo-lutive mixtures of speech using time-frequency mask and mixing matrix esti-mation,” IEICE Trans. Fundamentals, vol. E88-A, no. 7, pp. 1693-1700, 2005.CrossRefGoogle Scholar
  8. 8.
    K. Waheed and F. Salem, “Algebraic overcomplete independent component analysis,” in Proc. ICA 2003, 2003, pp. 1077-1082.Google Scholar
  9. 9.
    F. Theis,“Mathematics in independent component analysis,” Ph.D.dissertation, University of Regensburg,2002.[Online]. Available: http://homepages.uni-regensburg.de/ thf11669/phdthesis.html
  10. 10.
    A. Ferréol, L. Albera, and P. Chevalier, “Fourth-order blind identification of underdetermined mixtures of sources (FOBIUM),” IEEE Trans. on Signal Processing, vol. 53, no. 5, pp. 1640-1653, May 2005.CrossRefGoogle Scholar
  11. 11.
    L. D. Lathauwer and J. Castaing, “Second-order blind identification of underdetermined mixtures,” in 6th Int. Conference on Independent Component Analysis and Blind Signal Separation (ICA 2006), R. et al., Ed. Justinian ıncipe, and Simon Haykin Charleston, SC, USA: Springer, Mar. 2006, pp. 40-47. [Online]. Available: http://publi-etis.ensea.fr/2006/LC06
  12. 12.
    L. Albera, P. Comon, P. Chevalier, and A. Ferrol, “Blind identification of un-derdetermined mixtures based on the hexacovariance,” in Proc. ICASSP 2004, vol. II, May 2004, pp. 29-32.Google Scholar
  13. 13.
    P. Bofill and E. Monte, “Underdetermined convoluted source reconstruction using lp and socp, and a neural approximator of the optimizer,” in Indepen-dent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 569-576.Google Scholar
  14. 14.
    Y. Deville, J. Chappuis, S. Hosseini, and J. Thomas, “Differential fast fixed-point bss for underdetermined linear instantaneous mixtures,” in Indepen-dent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 48-56.Google Scholar
  15. 15.
    C. Wei, L. Khor, W. Woo, and S. Dlay, “Post-nonlinear underdetermined ICA by Bayesian statistics,” in Independent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 773-780.Google Scholar
  16. 16.
    S. Lesage, S. Krstulović, and R. Gribonval, “Under-determined source sep-aration: Comparison of two approaches based on sparse decompositions,” in Independent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 633-640.Google Scholar
  17. 17.
    C. Févotte and S. Godsill, “Blind separation of sparse sources using jeffrey’s inverse prior and the em algorithm,” in Independent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 593-600.Google Scholar
  18. 18.
    P. Comon and M. Rajih, “Blind identification of under-determined mixtures based on the characteristic function,” in ICASSP’05, vol. IV, Mar. 2005, pp. 1005-1008.Google Scholar
  19. 19.
    L. Albera, A. Ferreol, P. Comon, and P. Chevalier, “Blind Identification of Overcomplete MixturEs of sources (BIOME),” Linear Algebra Applications, Special Issue on Linear Algebra in Signal and Image Processing, vol. 391C, pp. 3-30, Nov. 2004.MathSciNetGoogle Scholar
  20. 20.
    L. D. Lathauwer, “Simultaneous matrix diagonalization: the overcomplete case,” in Proc. of the Fourth International Symposium on Independent Compo-nent Analysis and Blind Signal Separation (ICA 2003), Apr. 2003, pp. 821-825.Google Scholar
  21. 21.
    L. D. Lathauwer, B. D. Moor, J. Vandewalle, and J.-F. Cardoso, “Indepen-dent component analysis of largely underdetermined mixtures,” in Proc. of the Fourth International Symposium on Independent Component Analysis and Blind Signal Separation (ICA 2003), Apr. 2003, pp. 29-34.Google Scholar
  22. 22.
    L. Vielva, D. Erdogmus, C. Pantaleon, I. Santamaria, J. Pereda, and J. Principe, “Underdetermined blind source separation in a time-varying environment,” in Proc. ICASSP 2002, vol. 3, May 2002, pp. 3049-3052.Google Scholar
  23. 23.
    L. D. Lathauwer, P. Comon, B. D. Moor, and J. Vandewalle, “ICA algorithms for 3 sources and 2 sensors,” in Proc. IEEE Signal Processing Workshop on Higher-Order Statistics, Caesarea, Israel, 1999, pp. 116-120.Google Scholar
  24. 24.
    P. OGrady, B. Pearlmutter, and S. Rickard, “Survey of sparse and non-sparse methods in source separation,” International Journal of Imaging Systems and Technology, vol. 15, no. 1, pp. 18-33, July 2005.CrossRefGoogle Scholar
  25. 25.
    F. Abrard and Y. Deville, “A time-frequency blind signal separation method applicable to underdetermined mixtures of dependent sources,” Signal Processing, vol. 85, no. 7, pp. 1389-1403, July 2005. [Online]. Available: http://www.ast.obs-mip.fr/users/ydeville/papers/fa yd sigpro 2005 final%.pdf
  26. 26.
    N. Mitianoudis and T. Stathaki, “Overcomplete source separation using lapla-cian mixure models,” IEEE Signal Processing Letters, vol. 12, no. 4, pp. 277-280, Apr. 2005.CrossRefGoogle Scholar
  27. 27.
    S. Araki, H. Sawada, R. Mukai, and S. Makino, “A novel blind source separation method with observation vector clustering,” in Proc. IWAENC 2005, Sept. 2005, pp. 117-120.Google Scholar
  28. 28.
    R. Olsson and L. Hansen, “Blind separation of more sources than sensors in convolutive mixtures,” in Proc. ICASSP 2006, 2006.Google Scholar
  29. 29.
    M. Pedersen, D. Wang, J. Larsen, and U. Kjems, “Separating underdetermined convolutive speech mixtures,” in Independent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 674-681.Google Scholar
  30. 30.
    Y. Li, J. Wang, and A. Cichocki, “Blind source extraction from convolutive mixtures in ill-conditioned multi-input multi-output channels,” IEEE Trans. on Circuits and Systems - I: Regular Papers, vol. 51, no. 9, pp. 1814-1822, Sept. 2004.CrossRefMathSciNetGoogle Scholar
  31. 31.
    R. Saab, O. Yilmaz, M. McKeown, and R. Abugharbieh, “Underdetermined sparse blind source separation with delays,” in Signal Processing with Adaptive Sparse Structured Representations Workshop (SPARS), 2005.Google Scholar
  32. 32.
    M. Molla, K. Hirose, and N. Minematsu, “Separation of mixed audio signals by source localization and binary masking with hilbert spectrum,” in Indepen-dent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 641-648.Google Scholar
  33. 33.
    S. J. Godsill and C. Andrieu, “Bayesian separation and recovery of convolutively mixed autoregressive sources,” in Proc. ICASSP 1999, vol. III, 1999, pp. 1733-1736. [Online]. Available: citeseer.csail.mit.edu/349030.htmlGoogle Scholar
  34. 34.
    S. Winter, H. Sawada, S. Araki, and S. Makino, “Overcomplete BSS for con-volutive mixtures based on hierarchical clustering,” in Proc. ICA 2004, Sept. 2004, pp. 652-660.Google Scholar
  35. 35.
    S. Winter, H. Sawada, and S. Makino, “On real and complex valued L1-norm minimization for overcomplete blind source separation,” in 2005 IEEE Work-shop on Applications of Signal Processing to Audio and Acoustics (WASPAA), New Paltz, NY, USA, 2005, pp. 86-89.Google Scholar
  36. 36.
    L. Vielva, D. Erdogmus, and J. C. Principe, “Underdetermined blind source separation using a probabilistic source sparsity model,” in Proc. ICA 2001, 2001, pp. 675-679.Google Scholar
  37. 37.
    W. Kellermann and H. Buchner, “Wideband algorithms versus narrowband algorithms for adaptive filtering in the DFT domain,” in Proc. Asilomar Conf. on Signals, Systems, and Computers, vol. 2, Nov. 2003, pp. 1278-1282.Google Scholar
  38. 38.
    N. Linh-Trung, A. Belouchrani, K. Abed-Meraim, and B. Boashash, “Separat-ing more sources than sensors using time-frequency distributions,” EURASIP Journal on Applied Signal Processing, vol. 2005, no. 17, pp. 2828-2847, 2005.MATHCrossRefGoogle Scholar
  39. 39.
    H. Sawada, S. Araki, R. Mukai, and S. Makino, “Blind extraction of a dominant source signal from mixtures of many sources,” in Proc. ICASSP 2005, vol. III, 2005, pp. 61-64.Google Scholar
  40. 40.
    H. Sawada, R. Mukai, S. Araki, and S. Makino, “A robust and precise method for solving the permutation problem,” IEEE Trans. Speech and Audio Process-ing, vol. 12, pp. 530-538, Sept. 2004.CrossRefGoogle Scholar
  41. 41.
    K. Matsuoka, “Independent component analysis and its applications to sound signal separation,” in Proc. IWAENC 2003, Kyoto, Sept. 2003, pp. 15-18.Google Scholar
  42. 42.
    A. Jourjine, S. Rickard, and O. Yilmaz, “Blind separation of disjoint orthogonal signals: Demixing n sources from 2 mixtures,” Proc. ICASSP 2000, vol. 5, pp. 2985-2988, 2000.Google Scholar
  43. 43.
    M. Pedersen, T. Lehn-Schiøler, and J. Larsen, “BLUES from music: BLind Un-derdetermined Extraction of Sources from music,” in Independent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 392-399.Google Scholar
  44. 44.
    A. Mansour, M. Kawamoto, and C. Puntonet, “A time-frequency approach to blind separation of underdetermine mixture of sources,” in Proc. IASTED International Conference Applied Simuation and Modelling, Sept. 2003, pp. 413-418.Google Scholar
  45. 45.
    .M. Zibulevsky and B. Pearlmutter,“Blind source separation by sparse decomposition,” Neural Computations, vol.13, no.4, pp.863-882,2001. [Online]. Available: http://iew3.technion.ac.il/~mcib/
  46. 46.
    S. S. Chen, D. L. Donoho, and M. A. Saunders, “Atomic decomposition by basis pursuit,” Dept. Stat., Stanford Univ, Stanford, CA, Tech. Rep., 1995. [Online]. Available: http://www-stat.stanford.edu/~ donoho/Reports/1995/30401.pdf
  47. 47.
    P. Comon, “Blind channel identification and extraction of more sources than sensors,” in Proc. SPIE, 1998, pp. 2-13, keynote address.Google Scholar
  48. 48.
    A. Taleb, “An algorithm for the blind identication of N independent signal with 2 sensors,” in Proc. ISSPA 01, Aug. 2001, pp. 5-8.Google Scholar
  49. 49.
    J.-F. Cardoso, “Super-symmetric decomposition of the fourth-order cumulant tensor blind identification of more sources than sensors,” in Proc. ICASSP 91, vol. V, 1991, pp. 3109-3112.Google Scholar
  50. 50.
    L. Khor, W. Woo, and S. Dlay, “Non-sparse approach to underdetermined blind signal estimation,” in Proc. ICASSP 2005, 2005.Google Scholar
  51. 51.
    L. Benaroya, F. Bimbot, and R. Gribonval, “Audio source separation with a single sensor,” IEEE Trans. Audio, Speech and Language Processing, vol. 14, no. 1, pp. 191-199, Jan. 2006.CrossRefGoogle Scholar
  52. 52.
    T. Beierholm, B. Pedersen, and O. Winther, “Low complexity Bayesian single channel source separation,” in Proc. ICASSP 2003, 2003.Google Scholar
  53. 53.
    D. Ellis, “Prediction-driven computational auditory scene analysis,” Ph.D. dis-sertation, MIT, 1996.Google Scholar
  54. 54.
    J. Burred and T. Sikora, “On the use of auditory representations for sparsity-based sound source separation,” in Proc. IEEE Fifth Int. Conf. on Informa-tion, Communications and Signal Processing (ICICS), Bangkok, Thailand, Dec. 2005.Google Scholar
  55. 55.
    A. Hyvärinen, J. Karhunen, and E. Oja, Independent Component Analysis. New York: John Wiley & Sons, 2000.Google Scholar
  56. 56.
    F. Theis and E. Lang, “Formalization of the two-step approach to overcomplete BSS,” in Proc. of SIP 2002, Kauai, Hawaii, USA, 2002, pp. 207-212. [Online]. Available: http://homepages.uni-regensburg.de/~thf11669/ publications/theis02twostep SIP02.pdf
  57. 57.
    K. Waheed, “Blind source recovery: state space formulations,” Department of Electrical and Computer Engineering, Michigan State University, Tech. Rep., Sept. 2001.Google Scholar
  58. 58.
    P. Georgiev, P. G., D. Nuzillard, and A. Ralescu, “Sparse deflations in blind signal separation,” in Independent Component Analysis and Blind Signal Sep-aration, ser. LNCS, vol. 3889. Springer, 2006, pp. 807-814.Google Scholar
  59. 59.
    Y. Luo, W. Wang, J. Chambers, S. Lambotharan, and I. Proudler, “Exploita-tion of source nonstationarity in underdetermined blind source separation with advanced clustering techniques,” IEEE Trans. Signal Processing, vol. 54, no. 6, pp. 2198-2212, June 2006.CrossRefGoogle Scholar
  60. 60.
    C. Chang, P. C. Fung, and Y. S. Hung, “On a sparse component analysis approach to blind source separation,” in Independent Component Analysis and Blind Signal Separation, ser. LNCS, vol. 3889. Springer, 2006, pp. 765-772.Google Scholar
  61. 61.
    B. A. Pearlmutter and V. K. Potluru, “Sparse separation: Principles and tricks,” in Proc SPIE, vol. 5102, Apr. 2003, pp. 1-4.Google Scholar
  62. 62.
    I. Gorodnitsky and B. Rao, “Sparse signal reconstruction from limited data using FOCUSS: a re-weighted minimum norm algorithm,” IEEE Trans. Signal Processing, vol. 45, no. 3, pp. 600-616, Mar. 1997.CrossRefGoogle Scholar
  63. 63.
    T. Kristjansson, J. Hershey, and H. Attias, “Single microphone source separa-tion using high resolution signal reconstruction,” in Proc. ICASSP 2004, 2004.Google Scholar
  64. 64.
    A. Nesbit, M. Davies, M. Plumbley, and M. Sandler, “Source extraction from two-channel mixtures by joint cosine packet analysis,” in Proc. EUSICPO 2006, 2006.Google Scholar
  65. 65.
    L. D. Lathauwer, B. D. Moor, and J. Vandewalle, “Ica techniques for more sources than sensors,” in Proc. HOS 99, Caesarea, Israel, June 1999, pp. 121-124.Google Scholar
  66. 66.
    P. Comon and O. Grellier, “Non-linear inversion of underdetermined mixtures,” in Proc. ICA 99, 1999, pp. 461-465.Google Scholar
  67. 67.
    P. Comon, “Blind identification and source separation in 2x3 under-determined mixtures,” IEEE Trans. Signal Processing, vol. 52, no. 1, pp. 11-22, Jan. 2004.CrossRefMathSciNetGoogle Scholar
  68. 68.
    C. M. Bishop, Neural Networks for Pattern Recognition. Oxford University Press, 1995.Google Scholar
  69. 69.
    A. Gelman, J. Carlin, H. Stern, and D. Rubin, Bayesian Data Analysis. Chap-man & Hall, 1995.Google Scholar
  70. 70.
    D. Donoho and M. Elad,“Optimally-sparse representation in general (non-orthogonal) dictionaries via l1 minimization,” Proc. Nat. Aca. Sci, vol. 100, no.5, pp.2197-2202, Mar.2003.[Online]. Available: http://www.pnas.org/cgi/reprint/100/5/2197.pdf
  71. 71.
    T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learn-ing: Data Mining, Inference, and Prediction, ser. Springer Series in Statistics. Springer-Verlag, 2002.Google Scholar
  72. 72.
    F. Murtagh, “Comments on ‘Parallel algorithms for hierarchical clustering and cluster validity’,” IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 14, no. 10, pp. 1056-1057, Oct. 1992.CrossRefGoogle Scholar
  73. 73.
    A. Papoulis and S. Pillai, Probability, Random Variables, and Stochastic Processes, 4th ed. McGraw-Hill, 2002.Google Scholar
  74. 74.
    A. Pruessner, M. Bussieck, S. Dirkse, and A. Meeraus, “Conic programming in GAMS,” in INFORMS Annual Meeting, Atlanta, Oct. 2003, pp. 19-22. [Online]. Available: http://www.gams.com/presentations/present conic.pdf
  75. 75.
    J. Sturm, “Using SeDuMi 1.02, a MATLAB toolbox for optimization over symmetric cones,” Optimization Methods and Software, vol. 11-12, pp. 625-653, 1999, special issue on Interior Point Methods. [Online]. Available: http://fewcal.kub.nl/sturm/software/sedumi.html
  76. 76.
    L. S. Lobo, L. Vandenberghe, S. Boyd, and H. Lebert, “Second order cone programming,” Linear Algebra and Its Applications, vol. 284, pp. 193-228, 1998.MATHCrossRefMathSciNetGoogle Scholar
  77. 77.
    F. Alizadeh and D. Goldfarb, “Second-order cone programming,” Rugers Uni-versity, Tech. Rep., 2001.Google Scholar
  78. 78.
    S. Boyd and L. Vandenberghe, Convex Optimization. Cambridge University Press, 2004.Google Scholar
  79. 79.
    M. Lewicki and T. Sejnowski,“Learning overcomplete representations,” Neural Computation, vol. 12, no. 2, pp. 337-365, 2000. [Online]. Available: citeseer.nj.nec.com/lewicki98learning.htmlGoogle Scholar
  80. 80.
    I. Takigawa, M. Kudo, and J. Toyama, “Performance analysis of minimum ℓ1 -norm solutions for underdetermined source separation,” IEEE Trans. Signal Processing, vol. 52, no. 3, pp. 582-591, Mar. 2004.CrossRefMathSciNetGoogle Scholar
  81. 81.
    D. Malioutov, M. Cetin, and A. Willsky, “Optimal sparse representations in general overcomplete bases,” in Proc. ICASSP 2004, 2004, pp. 793-796.Google Scholar
  82. 82.
    E. Vincent, R. Gribonval, and C. Févotte, “Performance measurement in blind audio source separation,” IEEE Trans. Speech, Audio and Language Processing, vol. 14, no. 4, pp. 1462-1469, Jul. 2006.CrossRefGoogle Scholar
  83. 83.
    C. Févotte, R. Gribonval, and E. Vincent, “BSS EVAL toolbox user guide - Revision2.0,” IRISA, Tech. Rep.1706, Apr.2005.[Online]. Available: http://bass-db.gforge.inria.fr/bss eval/

Copyright information

© Springer 2007

Authors and Affiliations

  • Stefan Winter
    • 1
  • Walter Kellermann
    • 2
  • Hiroshi Sawada
    • 1
  • Shoji Makino
    • 1
  1. 1.NTT Communication Science LabsNTT CorporationSoraku-gunJapan
  2. 2.Multimedia Communications and Signal ProcessingUniversity of Erlangen-NurembergGermany

Personalised recommendations