Blind Audiovisual Source Separation Based on Redundant Representations

Anna Llagostera Casanovas 1 Gianluca Monaci 1 Pierre Vandergheynst 1 Rémi Gribonval 2
2 METISS - Speech and sound data modeling and processing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : In this work we present a method to perform a complete audiovisual source separation without need of previous information. This method is based on the assumption that sounds are caused by moving structures. Thus, an efficient representation of audio and video sequences allows to build relationships between synchronous structures on both modalities. A robust clustering algorithm groups video structures exhibiting strong correlations with the audio so that sources are counted and located in the image. Using such information and exploiting audio-video correlation, the audio sources activity is determined. Next, \backslashemph\char123spectral\char125 GMMs are learnt in time slots with only one source active so that it is possible to separate them in case of an audio mixture. Audio source separation performances are rigorously evaluated, clearly showing that the proposed algorithm performs efficiently and robustly.
Liste complète des métadonnées

Cited literature [10 references]  Display  Hide  Download

https://hal.inria.fr/inria-00544971
Contributor : Rémi Gribonval <>
Submitted on : Thursday, January 27, 2011 - 10:39:22 PM
Last modification on : Friday, November 16, 2018 - 1:23:46 AM
Document(s) archivé(s) le : Thursday, April 28, 2011 - 2:33:19 AM

File

2008_ICASSP_LlagosterasEtAl_BA...
Files produced by the author(s)

Identifiers

Citation

Anna Llagostera Casanovas, Gianluca Monaci, Pierre Vandergheynst, Rémi Gribonval. Blind Audiovisual Source Separation Based on Redundant Representations. Acoustics, Speech and Signal Processing, 2008. ICASSP 2008. IEEE International Conference on,, Apr 2008, Las Vegas, Nevada, United States. pp.1841 -1844, ⟨10.1109/ICASSP.2008.4517991⟩. ⟨inria-00544971⟩

Share

Metrics

Record views

394

Files downloads

198