Undercomplete Blind Subspace Deconvolution Via Linear Prediction

Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4701)


We present a novel solution technique for the blind subspace deconvolution (BSSD) problem, where temporal convolution of multidimensional hidden independent components is observed and the task is to uncover the hidden components using the observation only. We carry out this task for the undercomplete case (uBSSD): we reduce the original uBSSD task via linear prediction to independent subspace analysis (ISA), which we can solve. As it has been shown recently, applying temporal concatenation can also reduce uBSSD to ISA, but the associated ISA problem can easily become ‘high dimensional’ [1]. The new reduction method circumvents this dimensionality problem. We perform detailed studies on the efficiency of the proposed technique by means of numerical simulations. We have found several advantages: our method can achieve high quality estimations for smaller number of samples and it can cope with deeper temporal convolutions.


Independent Component Analysis Independent Component Analysis Linear Prediction Polynomial Matrix Blind Source Separation 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Szabó, Z., Póczos, B., Lőrincz, A.: Undercomplete blind subspace deconvolution. Journal of Machine Learning Research 8, 1063–1095 (2007)Google Scholar
  2. 2.
    Cichocki, A., Amari, S.: Adaptive blind signal and image processing. John Wiley & Sons, Chichester (2002)Google Scholar
  3. 3.
    Pedersen, M.S., Larsen, J., Kjems, U., Parra, L.C.: A survey of convolutive blind source separation methods. In: Springer Handbook of Speech, Springer, Heidelberg (to appear, 2007),
  4. 4.
    Cardoso, J.: Multidimensional independent component analysis. In: ICASSP 1998, vol. 4, pp. 1941–1944 (1998)Google Scholar
  5. 5.
    Rajagopal, R., Potter, L.C.: Multivariate MIMO FIR inverses. IEEE Transactions on Image Processing 12, 458–465 (2003)CrossRefMathSciNetGoogle Scholar
  6. 6.
    Févotte, C., Doncarli, C.: A unified presentation of blind source separation for convolutive mixtures using block-diagonalization. In: ICA 2003, pp. 349–354 (2003)Google Scholar
  7. 7.
    Choi, S., Cichocki, A.: Blind signal deconvolution by spatio-temporal decorrelation and demixing. Neural Networks for Signal Processing 7, 426–435 (1997)Google Scholar
  8. 8.
    Gorokhov, A., Loubaton, P.: Blind identification of MIMO-FIR systems: A generalized linear prediction approach. Signal Processing 73, 105–124 (1999)zbMATHCrossRefGoogle Scholar
  9. 9.
    Szabó, Z., Póczos, B., Lőrincz, A.: Cross-entropy optimization for independent process analysis. In: Rosca, J., Erdogmus, D., Príncipe, J.C., Haykin, S. (eds.) ICA 2006. LNCS, vol. 3889, pp. 909–916. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  10. 10.
    Theis, F.J.: Uniqueness of complex and multidimensional independent component analysis. Signal Processing 84, 951–956 (2004)CrossRefGoogle Scholar
  11. 11.
    Szabó, Z., Póczos, B., Lőrincz, A.: Undercomplete blind subspace deconvolution via linear prediction. Technical report, Eötvös Loránd University, Budapest (2007),
  12. 12.
    Theis, F.J.: Blind signal separation into groups of dependent signals using joint block diagonalization. In: ISCAS 2005, pp. 5878–5881 (2005)Google Scholar
  13. 13.
    Neumaier, A., Schneider, T.: Estimation of parameters and eigenmodes of multivariate AR models. ACM Trans. on Mathematical Software 27, 27–57 (2001)zbMATHCrossRefGoogle Scholar
  14. 14.
    Szabó, Z., Lőrincz, A.: Real and complex independent subspace analysis by generalized variance. In: ICARN 2006, pp. 85–88 (2006)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  1. 1.Department of Information Systems, Eötvös Loránd University, Pázmány P. sétány 1/C, Budapest H-1117Hungary

Personalised recommendations