Overcomplete ICA with a geometric algorithm

Fabian J. Theis, Elmar W. Lang, Tobias Westenhuber, Carlos G. Puntonet

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

We present an independent component analysis (ICA) algorithm based on geometric considerations [10] [11] to decompose a linear mixture of more sources than sensor signals. Bofill and Zibulevsky [2] recently proposed a two-step approach for the separation: first learn the mixing matrix, then recover the sources using a maximum-likelihood approach. We present an efficient method for the matrix-recovery step mimicking the standard geometric algorithm thus generalizing Bofill and Zibulevsky's method.

Original languageEnglish
Title of host publicationArtificial Neural Networks, ICANN 2002 - International Conference, Proceedings
EditorsJose R. Dorronsoro, Jose R. Dorronsoro
PublisherSpringer Verlag
Pages1049-1054
Number of pages6
ISBN (Print)9783540440741
DOIs
StatePublished - 2002
Externally publishedYes
Event2002 International Conference on Artificial Neural Networks, ICANN 2002 - Madrid, Spain
Duration: 28 Aug 200230 Aug 2002

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2415 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2002 International Conference on Artificial Neural Networks, ICANN 2002
Country/TerritorySpain
CityMadrid
Period28/08/0230/08/02

Fingerprint

Dive into the research topics of 'Overcomplete ICA with a geometric algorithm'. Together they form a unique fingerprint.

Cite this