ICA of correlated sources mismodeled as uncorrelated: Performance analysis

Dana Lahat; Hagit Messer; Jean François Cardoso

doi:10.1109/SSP.2009.5278531

ICA of correlated sources mismodeled as uncorrelated: Performance analysis

Dana Lahat, Hagit Messer, Jean François Cardoso

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

In this contribution we present closed-form performance analysis for the problem of blind separation of correlated sources when the separation algorithm is ignorant of the source model, and assumes only scalar sources. Our data model is of one scalar source and one multidimensional source. We obtain a closed-form analytical expression for applying joint diago-nalization on correlated sources. The obtained expression is different from the optimal expression (in the minimal mean square error (MMSE) sense), obtained by joint block diago-nalization. Numerical evaluations quantify the loss of performance due to mismodeling.

Original language	English
Title of host publication	2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09
Pages	489-492
Number of pages	4
DOIs	https://doi.org/10.1109/SSP.2009.5278531
State	Published - 2009
Externally published	Yes
Event	2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09 - Cardiff, United Kingdom Duration: 31 Aug 2009 → 3 Sep 2009

Publication series

Name	IEEE Workshop on Statistical Signal Processing Proceedings

Conference

Conference	2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09
Country/Territory	United Kingdom
City	Cardiff
Period	31/08/09 → 3/09/09

Keywords

Blind source separation
Independent component analysis
Joint diagonalization
Mismodeling

Access to Document

10.1109/SSP.2009.5278531

Cite this

@inproceedings{e744d098a87f451a8ca1cf0f0879875e,

title = "ICA of correlated sources mismodeled as uncorrelated: Performance analysis",

abstract = "In this contribution we present closed-form performance analysis for the problem of blind separation of correlated sources when the separation algorithm is ignorant of the source model, and assumes only scalar sources. Our data model is of one scalar source and one multidimensional source. We obtain a closed-form analytical expression for applying joint diago-nalization on correlated sources. The obtained expression is different from the optimal expression (in the minimal mean square error (MMSE) sense), obtained by joint block diago-nalization. Numerical evaluations quantify the loss of performance due to mismodeling.",

keywords = "Blind source separation, Independent component analysis, Joint diagonalization, Mismodeling",

author = "Dana Lahat and Hagit Messer and Cardoso, {Jean Fran{\c c}ois}",

year = "2009",

doi = "10.1109/SSP.2009.5278531",

language = "אנגלית",

isbn = "9781424427109",

series = "IEEE Workshop on Statistical Signal Processing Proceedings",

pages = "489--492",

booktitle = "2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09",

}

Lahat, D, Messer, H & Cardoso, JF 2009, ICA of correlated sources mismodeled as uncorrelated: Performance analysis. in 2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09., 5278531, IEEE Workshop on Statistical Signal Processing Proceedings, pp. 489-492, 2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09, Cardiff, United Kingdom, 31/08/09. https://doi.org/10.1109/SSP.2009.5278531

ICA of correlated sources mismodeled as uncorrelated: Performance analysis. / Lahat, Dana; Messer, Hagit; Cardoso, Jean François.
2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09. 2009. p. 489-492 5278531 (IEEE Workshop on Statistical Signal Processing Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - ICA of correlated sources mismodeled as uncorrelated

T2 - 2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09

AU - Lahat, Dana

AU - Messer, Hagit

AU - Cardoso, Jean François

PY - 2009

Y1 - 2009

N2 - In this contribution we present closed-form performance analysis for the problem of blind separation of correlated sources when the separation algorithm is ignorant of the source model, and assumes only scalar sources. Our data model is of one scalar source and one multidimensional source. We obtain a closed-form analytical expression for applying joint diago-nalization on correlated sources. The obtained expression is different from the optimal expression (in the minimal mean square error (MMSE) sense), obtained by joint block diago-nalization. Numerical evaluations quantify the loss of performance due to mismodeling.

AB - In this contribution we present closed-form performance analysis for the problem of blind separation of correlated sources when the separation algorithm is ignorant of the source model, and assumes only scalar sources. Our data model is of one scalar source and one multidimensional source. We obtain a closed-form analytical expression for applying joint diago-nalization on correlated sources. The obtained expression is different from the optimal expression (in the minimal mean square error (MMSE) sense), obtained by joint block diago-nalization. Numerical evaluations quantify the loss of performance due to mismodeling.

KW - Blind source separation

KW - Independent component analysis

KW - Joint diagonalization

KW - Mismodeling

UR - http://www.scopus.com/inward/record.url?scp=72349089982&partnerID=8YFLogxK

U2 - 10.1109/SSP.2009.5278531

DO - 10.1109/SSP.2009.5278531

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:72349089982

SN - 9781424427109

T3 - IEEE Workshop on Statistical Signal Processing Proceedings

SP - 489

EP - 492

BT - 2009 IEEE/SP 15th Workshop on Statistical Signal Processing, SSP '09

Y2 - 31 August 2009 through 3 September 2009

ER -

ICA of correlated sources mismodeled as uncorrelated: Performance analysis

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this