Citation: | CHENG Guanghui, MIAO Jifei, LI Wenrui, “Two Jacobi-like algorithms for the general joint diagonalization problem with applications to blind source separation,” Chinese Journal of Electronics, in press, doi: 10.23919/cje.2019.00.102, 2022. |
We consider the general problem of the approximate joint diagonalization of a set of non-Hermitian matrices. This problem mainly arises in the data model of the joint blind source separation for two datasets. Based on a special parameterization of the two diagonalizing matrices and on adapted approximations of the classical cost function, we establish two Jacobi-like algorithms. They may serve for the canonical polyadic decomposition (CPD) of a third-order tensor, and in some scenarios they can outperform traditional CPD methods. Simulation results demonstrate the competitive performance of the proposed algorithms.
[1] |
E. Moreau, “A generalization of joint diagonalization criteria for source separation”, IEEE Trans. Signal Processing, Vol.49, No.3, pp.530–541, 2001. doi: 10.1109/78.905873
|
[2] |
B. Afsari, “Simple LU and QR based non-orthogonal matrix joint diagonalization”, Independent Component Analysis and Blind Signal Separation, 6th International Conference, Charleston, SC, USA, pp. 1–7, 2006.
|
[3] |
P. Tichavsky and A. Yeredor, “Fast approximate joint diagonalization incorporating weight matrices”, IEEE Trans. Signal Processing, Vol.57, No.3, pp.878–891, 2009. doi: 10.1109/TSP.2008.2009271
|
[4] |
G. Chabriel, M. Kleinsteuber, E. Moreau, et al., “Joint matrices decompositions and blind source separation: A survey of methods, identification, and applications”, IEEE Signal Process. Mag., Vol.31, No.3, pp.34–43, 2014. doi: 10.1109/MSP.2014.2298045
|
[5] |
V. Maurandi and E. Moreau, “A decoupled Jacobi-like algorithm for non-unitary joint diagonalization of complex-valued matrices”, IEEE Signal Process. Lett., Vol.21, No.12, pp.1453–1456, 2014. doi: 10.1109/LSP.2014.2339891
|
[6] |
V. Maurandi, E. Moreau and C. DeLuigi, “Jacobi like algorithm for non-orthogonal joint diagonalization of Hermitian matrices”, ICASSP, Florence, Italy, pp. 6196–6200, 2014.
|
[7] |
T. Trainini and E. Moreau, “A coordinate descent algorithm for complex joint diagonalization under Hermitian and transpose congruences”, IEEE Trans. Signal Processing, Vol.62, No.19, pp.4974–4983, 2014. doi: 10.1109/TSP.2014.2343948
|
[8] |
G.H. Cheng, S.M. Li and E. Moreau, “New Jacobi-like algorithms for non-orthogonal joint diagonalization of Hermitian matrices”, Signal Processing, Vol.128, pp.440–448, 2016. doi: 10.1016/j.sigpro.2016.05.013
|
[9] |
D.Z. Feng, X.D. Zhang and Z. Bao, “An efficient multistage decomposition approach for independent components”, Signal Processing, Vol.83, No.1, pp.181–197, 2003. doi: 10.1016/S0165-1684(02)00390-0
|
[10] |
B. Afsari and P. S. Krishnaprasad, “Some gradient based joint diagonalization methods for ICA”, Independent Component Analysis and Blind Signal Separation, Fifth International Conference, Granada, Spain, pp. 437–444, 2004.
|
[11] |
H.J. Yu and D.S. Huang, “Graphical representation for DNA sequences via joint diagonalization of matrix pencil”, IEEE J. Biomedical and Health Informatics, Vol.17, No.3, pp.503–511, 2013. doi: 10.1109/TITB.2012.2227146
|
[12] |
J. F Cardoso and A. Souloumiac, “Blind beamforming for non-gaussian signals”, IEE Proceedings-F, Vol.140, No.6, pp.362–370, 1993.
|
[13] |
A. Belouchrani, K. Abed-Meraim, J.F. Cardoso and E. Moulines, “A blind source separation technique using second-order statistics”, IEEE Trans. Signal Processing, Vol.45, No.2, pp.434–444, 1997. doi: 10.1109/78.554307
|
[14] |
Y.O. Li, T. Adali, W. Wang and V. D. Calhoun, “Joint blind source separation by multiset canonical correlation analysis”, IEEE Trans. Signal Processing, Vol.57, No.10, pp.3918–3929, 2009. doi: 10.1109/TSP.2009.2021636
|
[15] |
V.D. Calhoun and T. Adali, “Multisubject independent component analysis of fmri: a decade of intrinsic networks, default mode, and neurodiagnostic discovery”, IEEE Reviews in Biomedical Engineering, Vol.5, pp.60–73, 2012. doi: 10.1109/RBME.2012.2211076
|
[16] |
V.D. Calhoun, T. Adali, G.D. Pearlson and J.J. Pekar, “A method for making group inferences from functional MRI data using independent component analysis”, Human Brain Mapping, Vol.14, No.3, Article No.140, 2001. doi: 10.1002/hbm.1048
|
[17] |
J. Laney, K. Westlake, S. Ma, E. Woytowicz, and T. Adali, “Capturing subject variability in data driven FMRI analysis: A graph theoretical comparison”, 48th Annual Conference on Information Sciences and Systems, Princeton, NJ, USA, pp. 1–6, 2014.
|
[18] |
A. M. Michael, M. Anderson, et al., “Preserving subject variability in group FMRI analysis: performance evaluation of GICA vs. IVA”, Frontiers in Systems Neuroscience, Vol.8, No.106, Article No.106, 2014.
|
[19] |
T. Adali, Y. Levin-Schwartz, and V. D. Calhoun, “Multimodal data fusion using source separation: Two effective models based on ICA and IVA and their properties”, Proceedings of the IEEE, Vol.103, No.9, pp.1478–1493, 2015. doi: 10.1109/JPROC.2015.2461624
|
[20] |
X.L. Li, T. Adali, and M. Anderson, “Joint blind source separation by generalized joint diagonalization of cumulant matrices”, Signal Processing, Vol.91, No.10, pp.2314–2322, 2011. doi: 10.1016/j.sigpro.2011.04.016
|
[21] |
P. Tichavsky, A. H. Phan, and A. Cichocki, “Non-orthogonal tensor diagonalization”, Signal Processing, Vol.138, pp.313–320, 2017. doi: 10.1016/j.sigpro.2017.04.001
|
[22] |
F. Roemer and M. Haardt, “A semialgebraic framework for approximate CP decompositions via simultaneous matrix diagonalizations”, Elsevier North-Holland, Inc., 2013.
|
[23] |
K. Naskovska, M. Haardt, P. Tichavsky, et al., “Extension of the semi-algebraic framework for approximate CP decompositions via non-symmetric simultaneous matrix diagonalization”, IEEE International Conference on Acoustics, Speech and Signal Processing, pp.2971–2975, 2016.
|
[24] |
A. Ziehe, P. Laskov, G. Nolte and K.R. Müller, “A fast algorithm for joint diagonalization with non-orthogonal transformations and its application to blind source separation”, Journal of Machine Learning Research, Vol.5, No.3, pp.777–800, 2004.
|
[25] |
O. Macchi and E. Moreau, “Self-adaptive source separation by direct and recursive networks”, International Conference on Digital Signal Processing, pp.1154–1159, 1154.
|
[26] |
A. Mesloub, K. Abed-Meraim and A. Belouchrani, “A new algorithm for complex nonorthogonal joint diagonalization based on Shear and Givens rotations”, IEEE Trans. Signal Processing, Vol.62, No.8, pp.1913–1925, 2014. doi: 10.1109/TSP.2014.2303947
|
[27] |
N. Vervliet, O. Debals, L. Sorber, et al., “Tensorlab 3.0”, 2016.
|
[28] |
M. Anderson, X.L. Li and T. Adali, “Complex-valued independent vector analysis: Application to multivariate Gaussian model”, Signal Processing, Vol.92, No.8, pp.1821–1831, 2012. doi: 10.1016/j.sigpro.2011.09.034
|
[29] |
L. Zou, X. Chen and Z. J. Wang, “Underdetermined joint blind source separation for two datasets based on tensor decomposition”, IEEE Signal Process. Lett., Vol.23, No.5, pp.673–677, 2016. doi: 10.1109/LSP.2016.2546687
|
[30] |
J. Benesty, J.D. Chen, Y.T. Huang, and I. Cohen, “Pearson correlation coefficient”, Springer Berlin Heidelberg, 2009.
|
[31] |
A. L. Goldberger, L. A. N. Amaral, L. Glass, et al., “Physiobank, physiotoolkit, and physionet components of a new research resource for complex physiologic signals”, Circulation, Vol.101, No.23, Article No.215, 2000.
|