Second-Order Blind Source Separation Based on Multi-dimensional Autocovariances

  • Fabian J. Theis
  • Anke Meyer-Bäse
  • Elmar W. Lang
Conference paper

DOI: 10.1007/978-3-540-30110-3_92

Part of the Lecture Notes in Computer Science book series (LNCS, volume 3195)
Cite this paper as:
Theis F.J., Meyer-Bäse A., Lang E.W. (2004) Second-Order Blind Source Separation Based on Multi-dimensional Autocovariances. In: Puntonet C.G., Prieto A. (eds) Independent Component Analysis and Blind Signal Separation. ICA 2004. Lecture Notes in Computer Science, vol 3195. Springer, Berlin, Heidelberg

Abstract

SOBI is a blind source separation algorithm based on time decorrelation. It uses multiple time autocovariance matrices, and performs joint diagonalization thus being more robust than previous time decorrelation algorithms such as AMUSE. We propose an extensioncalled mdSOBI by using multidimensional autocovariances, which can be calculated for data sets with multidimensional parameterizations such as images or fMRI scans. mdSOBI has the advantage of using the spatial data in all directions, whereas SOBI only uses a single direction. These findings are confirmed by simulations and an application to fMRI analysis, where mdSOBI outperforms SOBI considerably.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Fabian J. Theis
    • 1
    • 2
  • Anke Meyer-Bäse
    • 2
  • Elmar W. Lang
    • 1
  1. 1.Institute of BiophysicsUniversity of RegensburgRegensburgGermany
  2. 2.Department of Electrical and Computer EngineeringFlorida State UniversityTallahasseeUSA

Personalised recommendations