Quadratic MIMO contrast functions for blind source separation in a convolutive context

Saloua Rhioui, Marc Castella, Eric Moreau

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

Abstract

This paper considers the problem of blind separation of sources mixed by a MIMO convolutive system. For both i.i.d. and non i.i.d. sources, quadratic separation criteria previously designed for the extraction of a single source are extended to parallel extraction in the MIMO case. These criteria are based on the use of so-called reference signals and a condition is given under which we obtain MIMO contrast functions. Simulations demonstrate that a particular choice of a set of reference signals ensures the contrast property. The performance offered by these criteria is investigated through simulations: it is shown that the proposed contrast functions avoid accumulation errors, contrary to deflation methods.

Original languageEnglish
Title of host publicationIndependent Component Analysis and Blind Signal Separation - 6th International Conference, ICA 2006, Proceedings
Pages230-237
Number of pages8
DOIs
Publication statusPublished - 11 Jul 2006
Event6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006 - Charleston, SC, United States
Duration: 5 Mar 20068 Mar 2006

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume3889 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006
Country/TerritoryUnited States
CityCharleston, SC
Period5/03/068/03/06

Fingerprint

Dive into the research topics of 'Quadratic MIMO contrast functions for blind source separation in a convolutive context'. Together they form a unique fingerprint.

Cite this