Blind identification problems with constraints

A. Cichocki1, P. Georgiev1
1Laboratory for Advanced Brain Signal Processing, Brain Science Institute, RIKEN Institute of Physical and Chemical Research, Wako, Saitama, Japan

Tóm tắt

In many applications of independent component analysis (ICA) and blind source separation (BSS) the mixing or separating matrices have some special structure or some constraints are imposed for the matrices like symmetry, orthogonality, nonnegativity, sparseness and unit (or specified invariant norm) of the matrix. We present several algorithms and overview some known transformations which allows us to preserve such constraints. Especially, we propose algorithms for a blind identification problem with non-negativity constraints.

Từ khóa

#Independent component analysis #Symmetric matrices #Source separation #Covariance matrix #Blind source separation #Signal processing algorithms #Infrared sensors #Matrix decomposition #Laboratories #Biomedical signal processing

Tài liệu tham khảo

10.1088/0266-5611/16/5/316 10.1002/0471221317 10.1109/78.845952 nagy, 2000, Enforcing non-negativity in image reconstruction algorithms, Mathematical Modeling Estimation and Imaging, 182, 10.1117/12.402439 lee, 1999, Learning of the parts of objects by non-negative matrix factorization, Nature, 401, 10.1038/44565 10.1137/S0895479895290954 10.1109/NNSP.2002.1030067 10.1109/ICASSP.2001.940338 plumbley, 0, Adaptive lateral inhibition for non-negative ICA, Proc ICA?2001 georgiev, 2001, On some extensions of the natural gradient algorithm, Proc Third Int Conf Independent Component Analysis and Blind Signal Separation, 581 calhoun, 2001, Group ICA of functional MRI data: separability, stationarity, and inference, Proc ICA 10.1002/0470845899