Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Blind Source SeparationDependent Component Analysis Exploiting Nonnegativity and/or Time-Domain Sparsity

Blind Source Separation: Dependent Component Analysis Exploiting Nonnegativity and/or Time-Domain... [It is well-known that many real-world signals are nonnegative [1–8], i.e., their sample values are either zero or greater than zero, such as images. Obviously, nonnegativity is different from the statistical information of sources. Depending on the kinds of dependent sources, the nonnegativity of the source signals could be exploited to carry out dependent component analysis (DCA), i.e., separate these unknown dependent sources from their observed mixtures. If the sources also have certain level of sparsity in time domain, then the nonnegativity and time-domain sparsity of the source signals can be jointly employed to achieve DCA. In this chapter, three classes of dependent component analysis methods are introduced and analyzed, which are the nonnegative sparse representation (NSR) based methods, the convex geometry analysis (CGA) based methods, and the nonnegative matrix factorization (NMF) based methods. These methods either exploit the nonnegativity of the sources or both the nonnegativity and time-domain sparsity of the sources.] http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png

Blind Source SeparationDependent Component Analysis Exploiting Nonnegativity and/or Time-Domain Sparsity

Download PDF
28 pages

Loading next page...
 
/lp/springer-journals/blind-source-separation-dependent-component-analysis-exploiting-GN04ZdhGNl
Publisher
Springer Singapore
Copyright
© The Author(s) 2015
ISBN
978-981-287-226-5
Pages
19 –47
DOI
10.1007/978-981-287-227-2_2
Publisher site
See Chapter on Publisher Site

Abstract

[It is well-known that many real-world signals are nonnegative [1–8], i.e., their sample values are either zero or greater than zero, such as images. Obviously, nonnegativity is different from the statistical information of sources. Depending on the kinds of dependent sources, the nonnegativity of the source signals could be exploited to carry out dependent component analysis (DCA), i.e., separate these unknown dependent sources from their observed mixtures. If the sources also have certain level of sparsity in time domain, then the nonnegativity and time-domain sparsity of the source signals can be jointly employed to achieve DCA. In this chapter, three classes of dependent component analysis methods are introduced and analyzed, which are the nonnegative sparse representation (NSR) based methods, the convex geometry analysis (CGA) based methods, and the nonnegative matrix factorization (NMF) based methods. These methods either exploit the nonnegativity of the sources or both the nonnegativity and time-domain sparsity of the sources.]

Published: Sep 17, 2014

Keywords: Nonnegative matrix factorization; Sparse representation; Convex geometry analysis

There are no references for this article.