Stability of independent vector analysis

Independent vector analysis (IVA) is a method for solving the permutation problem that is inherent in the frequency-domain independent component analysis for convoluted mixtures. IVA utilizes inner dependency among the frequency components of each source. It is formulated as a problem of minimizing a measure that represents difference between a prior probability density function (pdf) of the set of frequency components and the actual pdf of the output of the demixing process. Although the effectiveness of the IVA method has been demonstrated by many applications, there are very few mathematical analyses of the algorithm. This paper shows a remarkable proposition that proves the validity of IVA: if the desired demixing process minimizes the measure, any permuted one never becomes a local minimum.

[1]  Kenneth Kreutz-Delgado,et al.  Probabilistic Formulation of Independent Vector Analysis Using Complex Gaussian Scale Mixtures , 2009, ICA.

[2]  Nobuhiko Kitawaki,et al.  Combined approach of array processing and independent component analysis for blind separation of acoustic signals , 2003, IEEE Trans. Speech Audio Process..

[3]  Tülay Adali,et al.  Nonorthogonal Independent Vector Analysis Using Multivariate Gaussian Model , 2010, LVA/ICA.

[4]  Te-Won Lee,et al.  Fast fixed-point independent vector analysis algorithms for convolutive blind source separation , 2007, Signal Process..

[5]  Te-Won Lee,et al.  Independent Vector Analysis using Non-Spherical Joint Densities for the Separation of Speech Signals , 2007, 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07.

[6]  Lucas C. Parra,et al.  Convolutive blind separation of non-stationary sources , 2000, IEEE Trans. Speech Audio Process..

[7]  Taesu Kim,et al.  Real-Time Independent Vector Analysis for Convolutive Blind Source Separation , 2010, IEEE Transactions on Circuits and Systems I: Regular Papers.

[8]  Kazuya Takeda,et al.  Evaluation of blind signal separation method using directivity pattern under reverberant conditions , 2000, 2000 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings (Cat. No.00CH37100).

[9]  Te-Won Lee,et al.  On the Assumption of Spherical Symmetry and Sparseness for the Frequency-Domain Speech Model , 2007, IEEE Transactions on Audio, Speech, and Language Processing.

[10]  Andreas Ziehe,et al.  An approach to blind source separation based on temporal structure of speech signals , 2001, Neurocomputing.

[11]  Te-Won Lee,et al.  Blind Source Separation Exploiting Higher-Order Frequency Dependencies , 2007, IEEE Transactions on Audio, Speech, and Language Processing.

[12]  Pierre Comon,et al.  Independent component analysis, A new concept? , 1994, Signal Process..

[13]  Atsuo Hiroe,et al.  Solution of Permutation Problem in Frequency Domain ICA, Using Multivariate Probability Density Functions , 2006, ICA.

[14]  E. Oja,et al.  Independent Component Analysis , 2013 .

[15]  Hualiang Li,et al.  Algorithms for Complex ML ICA and Their Stability Analysis Using Wirtinger Calculus , 2010, IEEE Transactions on Signal Processing.

[16]  Hiroshi Sawada,et al.  A robust and precise method for solving the permutation problem of frequency-domain blind source separation , 2004, IEEE Transactions on Speech and Audio Processing.

[17]  K. Matsuoka,et al.  Stability analysis on independent vector analysis , 2011 .

[18]  Paris Smaragdis,et al.  Blind separation of convolved mixtures in the frequency domain , 1998, Neurocomputing.

[19]  Hiroshi Sawada,et al.  Spatio–Temporal FastICA Algorithms for the Blind Separation of Convolutive Mixtures , 2007, IEEE Transactions on Audio, Speech, and Language Processing.

[20]  Terrence J. Sejnowski,et al.  Independent Vector Analysis for Source Separation Using a Mixture of Gaussians Prior , 2010, Neural Computation.