Sparse component analysis and blind source separation of underdetermined mixtures

被引:265
作者
Georgiev, P [1 ]
Theis, F
Cichocki, A
机构
[1] Univ Cincinnati, ECECS Dept, Cincinnati, OH 45221 USA
[2] Univ Regensburg, Inst Biophys, D-93040 Regensburg, Germany
来源
IEEE TRANSACTIONS ON NEURAL NETWORKS | 2005年 / 16卷 / 04期
关键词
blind source separation (BSS); sparse component analysis (SCA); underdetermined mixtures;
D O I
10.1109/TNN.2005.849840
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In this letter, we solve the problem of identifying matrices S is an element of R-n x N and A is an element of R-m x n knowing only their multiplication X = AS, under some conditions, expressed either in terms of A and sparsity of S (identifiability conditions), or in terms of X (sparse component analysis (SCA) conditions). We present algorithms for such identification and illustrate them by examples.
引用
收藏
页码:992 / 996
页数:5
相关论文
共 11 条
[1]  
ABRARD F, 2001, P 3 INT C IND COMP A, P734
[2]  
[Anonymous], 2002, Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications
[3]  
Chen Y, 1998, NONCON OPTIM ITS APP, V20, P1
[4]   Optimally sparse representation in general (nonorthogonal) dictionaries via l1 minimization [J].
Donoho, DL ;
Elad, M .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2003, 100 (05) :2197-2202
[5]  
Hyvärinen A, 2001, INDEPENDENT COMPONENT ANALYSIS: PRINCIPLES AND PRACTICE, P71
[6]  
Jourjine A, 2000, INT CONF ACOUST SPEE, P2985, DOI 10.1109/ICASSP.2000.861162
[7]   Learning the parts of objects by non-negative matrix factorization [J].
Lee, DD ;
Seung, HS .
NATURE, 1999, 401 (6755) :788-791
[8]   Blind source separation of more sources than mixtures using overcomplete representations [J].
Lee, TW ;
Lewicki, MS ;
Girolami, M ;
Sejnowski, TJ .
IEEE SIGNAL PROCESSING LETTERS, 1999, 6 (04) :87-90
[9]   A geometric algorithm for overcomplete linear ICA [J].
Theis, TJ ;
Lang, EW ;
Puntonet, CG .
NEUROCOMPUTING, 2004, 56 :381-398
[10]  
Waheed K, 2003, P 4 INT S IND COMP A, P1077