Journal of Guangdong University of Technology ›› 2008, Vol. 25 ›› Issue (4): 61-64.

• Comprehensive Studies • Previous Articles     Next Articles

Underdetermined Blind Source Separation Based on Locally Sparse Representations

  

  1. 1.Faculty of Applied Mathematics,Guangdong University of Technology,Guangzhou 510006,China;2.Faculty of Information Engineering,JiangXi University of Science and Technology,Ganzhou 341000
  • Online:2008-03-01 Published:2008-03-01

HTML

PDF (PC)

2144

Abstract: It proposes a new algorithm of locally sparse blind source separation for estimating the mixed matrix A.This method doesn’t require all the samples of the sources are strictly sparse.Firstly all the vectors in the same line were searched for and classified.And then the mixed matrix A determined.The simulation illustrates the effectiveness of this algorithrn.

Key words: blind seurce separation； sparse； strictly sparse； underdetermined；

[1] 肖明,谢胜利,傅予力.  欠定情形下语音信号盲分离的时域检索平均法[J]. 中国科学(E辑:信息科学). 2007(12)

[2] 谢胜利,谭北海,傅予力.  基于平面聚类算法的欠定混叠盲信号分离[J]. 自然科学进展. 2007(06)

[3] 何昭水,谢胜利,傅予力.  稀疏表示与病态混叠盲分离[J]. 中国科学E辑:信息科学. 2006(08)

[4] BELOUCHRANI A,CARDOSO J F.Maximum likelihoodsource separation for discrete sources. Proc.EUSIP-CO . 1994

[5] He Z.S,,Cichocki A.K-EVD clustering and its applications to sparse component analysis. Proc. of the 6th International Conference on Independent Component Analysis and Blind Signal Separation . 2006

[6] Georgiev P,Theis F,Cichocki A.Sparse component analysis and blind source separation of underdetermined mixtures. IEEE Transactions of Neural Networks . 2005

[7] Li Y,Andrzej C,Amari S.Analysis of sparse representation and blind source separation. Neural Computation . 2004
No related articles found!
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
Copyright © 2017 Journal of Guangdong University of Technology
Add：729 East Dongfeng RD., Guangzhou PRC. Postcode: 510090
Tel：020-37626653,020-37626139,020-37626396
Email：xbzrb@gdut.edu.cn
Supported:Beijing Magtech