A New Link Between Joint Blind Source Separation Using Second Order Statistics and the Canonical Polyadic Decomposition

In this paper, we discuss the joint blind source separation (JBSS) of real-valued Gaussian stationary sources with uncorrelated samples from a new perspective. We show that the second-order statistics of the observations can be reformulated as a coupled decomposition of several tensors. The canonical polyadic decomposition (CPD) of each such tensor, if unique, results in the identification of one or two mixing matrices. The proposed new formulation implies that standard algorithms for joint diagonalization and CPD may be used to estimate the mixing matrices, although only in a sub-optimal manner. We discuss the uniqueness and identifiability of this new approach. We demonstrate how the proposed approach can bring new insights on the uniqueness of JBSS in the presence of underdetermined mixtures.

[1]  Tülay Adali,et al.  A maximum likelihood approach for independent vector analysis of Gaussian data sets , 2011, 2011 IEEE International Workshop on Machine Learning for Signal Processing.

[2]  T. W. Anderson,et al.  An Introduction to Multivariate Statistical Analysis , 1959 .

[3]  Ronald Phlypo,et al.  Independent Vector Analysis: Identification Conditions and Performance Bounds , 2013, IEEE Transactions on Signal Processing.

[4]  Christian Jutten,et al.  An alternative proof for the identifiability of independent vector analysis using second order statistics , 2016, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[5]  Christian Jutten,et al.  Joint Independent Subspace Analysis Using Second-Order Statistics , 2016, IEEE Transactions on Signal Processing.

[6]  Christian Jutten,et al.  Joint analysis of multiple datasets by cross-cumulant tensor (block) diagonalization , 2016, 2016 IEEE Sensor Array and Multichannel Signal Processing Workshop (SAM).

[7]  J. Kruskal.  Three-way arrays: rank and uniqueness of trilinear decompositions, with application to arithmetic complexity and statistics , 1977 .

[8]  Qiu-Hua Lin,et al.  Double Coupled Canonical Polyadic Decomposition for Joint Blind Source Separation , 2016, IEEE Transactions on Signal Processing.

[9]  Te-Won Lee,et al.  Independent Vector Analysis: An Extension of ICA to Multivariate Components , 2006, ICA.

[10]  Christian Jutten,et al.  Multimodal Data Fusion: An Overview of Methods, Challenges, and Prospects , 2015, Proceedings of the IEEE.

[11]  Ronald Phlypo,et al.  Orthogonal and non-orthogonal joint blind source separation in the least-squares sense , 2012, 2012 Proceedings of the 20th European Signal Processing Conference (EUSIPCO).

[12]  Tülay Adali,et al.  Joint blind source separation by generalized joint diagonalization of cumulant matrices , 2011, Signal Process..

[13]  Vince D. Calhoun,et al.  Joint Blind Source Separation by Multiset Canonical Correlation Analysis , 2009, IEEE Transactions on Signal Processing.

[14]  Tülay Adali,et al.  Joint blind source separation from second-order statistics: Necessary and sufficient identifiability conditions , 2011, 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[15]  F. L. Hitchcock.  The Expression of a Tensor or a Polyadic as a Sum of Products , 1927 .