Methods of Fair Comparison of Performance of Linear ICA Techniques in Presence of Additive Noise

Linear ICA model with additive Gaussian noise is frequently considered in many practical applications, because it approaches the reality often much better than the noise-free alternate. In this paper, a number important differences between noisy and noiseless ICA are discussed. It is shown that estimation of the mixing/demixing matrix should not be the main goal, in the noisy case. Instead, it is proposed to compare the outcome of ICA algorithms with a minimum mean square (MMSE) separation, derived for known mixing model. The signal-to-interference-plus-noise ratio is suggested as the most meaningful performance criterion. A simulation study that compares a few well known ICA algorithms applied to noise data is included

[1]  E. Oja,et al.  Independent Component Analysis , 2001 .

[2]  V. Koivunen,et al.  Identifiability and Separability of Linear Ica Models Revisited , 2003 .

[3]  P. Tichavský,et al.  Efficient variant of algorithm fastica for independent component analysis attaining the cramer-RAO lower bound , 2005, IEEE/SP 13th Workshop on Statistical Signal Processing, 2005.

[4]  Alexander M. Bronstein,et al.  Relative optimization for blind deconvolution , 2005, IEEE Transactions on Signal Processing.

[5]  D. Chakrabarti,et al.  A fast fixed - point algorithm for independent component analysis , 1997 .

[6]  Daniel Yellin,et al.  Multichannel system identification and deconvolution: performance bounds , 1999, IEEE Trans. Signal Process..

[7]  J. Cardoso.  On the Performance of Orthogonal Source Separation Algorithms , 1994 .

[8]  Eric Moulines,et al.  Maximum likelihood for blind separation and deconvolution of noisy signals using mixture models , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[9]  Andrzej Cichocki,et al.  Bias removal technique for blind source separation with noisy measurements , 1998 .

[10]  Russell H. Lambert,et al.  OVERDETERMINED BLIND SOURCE SEPARATION: USING MORE SENSORS THAN SOURCE SIGNALS IN A NOISY MIXTURE , 2000 .

[11]  Tzyy-Ping Jung,et al.  Independent Component Analysis of Electroencephalographic Data , 1995, NIPS.

[12]  M. Davies,et al.  Identifiability issues in noisy ICA , 2004, IEEE Signal Processing Letters.

[13]  E. Oja,et al.  Corrections to "Performance Analysis of the FastICA Algorithm and Cram´ er-Rao Bounds for Linear Independent Component Analysis", TSP 04/06 , 2007 .

[14]  A. Yeredor.  Blind separation of Gaussian sources via second-order statistics with asymptotically optimal weighting , 2000, IEEE Signal Processing Letters.

[15]  Aapo Hyvärinen,et al.  Gaussian moments for noisy independent component analysis , 1999, IEEE Signal Processing Letters.

[16]  Erkki Oja,et al.  Cramer-Rao lower bound for linear independent component analysis , 2005, Proceedings. (ICASSP '05). IEEE International Conference on Acoustics, Speech, and Signal Processing, 2005..

[17]  Jean-Franois Cardoso.  High-Order Contrasts for Independent Component Analysis , 1999, Neural Computation.

[18]  Erkki Oja,et al.  Performance analysis of the FastICA algorithm and Crame/spl acute/r-rao bounds for linear independent component analysis , 2006, IEEE Transactions on Signal Processing.

[19]  Philippe Garat,et al.  Blind separation of mixture of independent sources through a quasi-maximum likelihood approach , 1997, IEEE Trans. Signal Process..

[20]  Ehud Weinstein,et al.  Maximum likelihood and lower bounds in system identification with non-Gaussian inputs , 1994, IEEE Trans. Inf. Theory.

[21]  T. Ens,et al.  Blind signal separation : statistical principles , 1998 .

[22]  Jean-François Cardoso,et al.  Equivariant adaptive source separation , 1996, IEEE Trans. Signal Process..

[23]  Erkki Oja,et al.  Efficient Variant of Algorithm FastICA for Independent Component Analysis Attaining the CramÉr-Rao Lower Bound , 2006, IEEE Transactions on Neural Networks.

[24]  S.C. Douglas,et al.  On the relationship between blind deconvolution and blind source separation , 1997, Conference Record of the Thirty-First Asilomar Conference on Signals, Systems and Computers (Cat. No.97CB36136).