Temporal Decorrelation as Preprocessing for Linear and Post-nonlinear ICA

We present a straightforward way to use temporal decorrelation as preprocessing in linear and post-nonlinear independent component analysis (ICA) with higher order statistics (HOS). Contrary to the separation methods using second order statistics (SOS), the proposed method can be applied when the sources have similar temporal structure. The main idea is that componentwise decorrelation increases non-Gaussianity and therefore makes it easier to separate sources with HOS ICA. Conceptually, the non-Gaussianizing filtering matches very well with the Gaussianization used to cancel the post-nonlinear distortions. Examples demonstrating the consistent improvement in the separation quality are provided for the both linear and post-linear cases.

[1]  Kiyohiro Shikano,et al.  Stable Learning Algorithm for Blind Separation of Temporally Correlated Acoustic Signals Combining Multistage ICA and Linear Prediction , 2003, IEICE Trans. Fundam. Electron. Commun. Comput. Sci..

[2]  L. Jackson.  Digital filters and signal processing , 1985 .

[3]  J. Cardoso,et al.  Blind beamforming for non-gaussian signals , 1993 .

[4]  M. Kendall,et al.  Kendall's advanced theory of statistics , 1995 .

[5]  M. Kendall,et al.  Kendall's Advanced Theory of Statistics: Volume 1 Distribution Theory , 1987 .

[6]  Asoke K. Nandi,et al.  BLIND SEPARATION OF ACOUSTIC MIXTURES BASED ON LINEAR PREDICTION ANALYSIS , 2003 .

[7]  Toshihisa Tanaka,et al.  Subband decomposition independent component analysis and new performance criteria , 2004, 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[8]  Visa Koivunen,et al.  Blind separation methods based on Pearson system and its extensions , 2002, Signal Process..

[9]  J. Karhunen,et al.  Advances in Nonlinear Blind Source Separation , 2003 .

[10]  Motoaki Kawanabe,et al.  Blind Separation of Post-nonlinear Mixtures using Linearizing Transformations and Temporal Decorrelation , 2003, J. Mach. Learn. Res..

[11]  Andreas Ziehe,et al.  Combining time-delayed decorrelation and ICA: towards solving the cocktail party problem , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[12]  Eric Moulines,et al.  A blind source separation technique using second-order statistics , 1997, IEEE Trans. Signal Process..

[13]  Aapo Hyvärinen,et al.  Independent Component Analysis for Time-dependent Stochastic Processes , 1998 .

[14]  Andrzej Cichocki,et al.  Blind signal extraction of signals with specified frequency band , 2002, Proceedings of the 12th IEEE Workshop on Neural Networks for Signal Processing.

[15]  A. Kaiser,et al.  CONSIDERING TEMPORAL STRUCTURES IN INDEPENDENT COMPONENT ANALYSIS , 2003 .

[16]  Andrzej Cichocki,et al.  Adaptive Blind Signal and Image Processing - Learning Algorithms and Applications , 2002 .

[17]  Danilo P. Mandic,et al.  AN ONLINE ALGORITHM FOR BLIND EXTRACTION OF SOURCES WITH DIFFERENT DYNAMICAL STRUCTURES , 2003 .