Une extension avec variables latentes du modèle d'analyse en composantes indépendantes

Abstract : This paper deals with the problem of blind source separation of instantaneous linear mixtures. The Independent Component Analysis (ICA) model is extended to a new model where the sources can be dependent. Depending on the value of a latent process, the unknown sources are assumed either mutually independent or dependent. We propose a separation method which combines: (i) a classical ICA separation performed using the set of samples whose components are conditionally independent, and (ii) a method for estimation of the latent process using the Iterative Conditional Estimation (ICE) method. Simulations confirm the great generality of our model and show that our method takes into account a partial dependence between the sources.
Document type :
Conference papers
Liste complète des métadonnées

Cited literature [7 references]  Display  Hide  Download

Contributor : Médiathèque Télécom Sudparis & Institut Mines-Télécom Business School <>
Submitted on : Monday, February 13, 2017 - 9:55:03 AM
Last modification on : Thursday, February 7, 2019 - 2:24:28 PM
Document(s) archivé(s) le : Sunday, May 14, 2017 - 12:25:08 PM


Files produced by the author(s)


  • HAL Id : hal-01465735, version 1


Selwa Rafi, Marc Castella, Wojciech Pieczynski. Une extension avec variables latentes du modèle d'analyse en composantes indépendantes. GRETSI 2011 : XXIIIe Colloque, Sep 2011, Bordeaux, France. pp.1 - 4. ⟨hal-01465735⟩



Record views


Files downloads