Bertrand Rivet
GIPSA-Lab, CNRS-UMR 5216, Grenoble INP
Session JEP poster P1 Lundi 9 Juin - 16h00 18h00
-
papier 1612
Séparation de sources non-stationnaires par la parcimonie pour un mélange linéaire instantané
- Bertrand Rivet ( GIPSA-Lab, CNRS-UMR 5216, Grenoble INP)
- Résumé : We propose a method to estimate non-stationnary sources with non activity periods in a determined linear instantaneous mixture. Our method is based on the assumption that in some unknown temporal periods speech signals are inactive leading thus to an overdetermined mixture. Such silence periods allow to estimate the rows of the demixing matrix by a new algorithm called Direction Estimation of Separating Matrix (DESM). The periods of sources inactivity are estimated by a generalized eigen decomposition of covariance matrices of the mixtures, and the separating matrix is then estimated by a kernel principal component analysis. Experiments are provided with determined mixtures, and shown to be efficient.
- article