Information theoretic versus cumulant-based contrasts for multimodal source separation

Vrins, Frédéric;Verleysen, Michel
(2005) IEEE Signal Processing Letters — Vol. 12, n° 3, p. 190-193 (2005)

Files

J28-Informationtheoreticversuscumulant-basedcontrastsformultimodalsourceseparation.pdf
  • Restricted Access
  • Adobe PDF
  • 190.51 KB
Request a copy

Details

Authors
Abstract
Recently, several authors have emphasized the existence of spurious maxima in usual contrast functions for source separation (e.g., the likelihood and the mutual information) when several sources have multimodal distributions. The aim of this letter is to compare the information theoretic contrasts to cumulant-based ones from the robustness to spurious maxima point of view. Even if all of them tend to measure, in some way, the same quantity, which is the output independence (or equivalently, the output non-Gaussianity), it is shown that in the case of a mixture involving two sources, the kurtosis-based contrast functions are more robust than the information theoretic ones when the source distributions are multimodal.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Vrins, F., & Verleysen, M. (2005). Information theoretic versus cumulant-based contrasts for multimodal source separation. IEEE Signal Processing Letters, 12(3), 190-193. https://doi.org/10.1109/LSP.2004.840863 (Original work published 2005)