Skip to Main content Skip to Navigation
Conference papers

Itakura-Saito nonnegative matrix factorization with group sparsity

Augustin Lefevre 1, 2, 3 Francis Bach 1, 3 Cédric Févotte 2
3 SIERRA - Statistical Machine Learning and Parsimony
DI-ENS - Département d'informatique de l'École normale supérieure, Inria Paris-Rocquencourt, CNRS - Centre National de la Recherche Scientifique : UMR8548
Abstract : We propose an unsupervised inference procedure for audio source separation. Components in nonnegative matrix factorization (NMF) are grouped automatically in audio sources via a penalized maximum likelihood approach. The penalty term we introduce favors sparsity at the group level, and is motivated by the assumption that the local amplitude of the sources are independent. Our algorithm extends multiplicative updates for NMF; moreover we propose a test statistic to tune hyperparameters in our model, and illustrate its adequacy on synthetic data. Results on real audio tracks show that our sparsity prior allows to identify audio sources without knowledge on their spectral properties.
Complete list of metadatas

Cited literature [7 references]  Display  Hide  Download
Contributor : Francis Bach <>
Submitted on : Sunday, February 20, 2011 - 10:06:29 PM
Last modification on : Friday, February 12, 2021 - 9:22:04 AM
Long-term archiving on: : Tuesday, November 6, 2012 - 2:30:27 PM


Publisher files allowed on an open archive


  • HAL Id : hal-00567344, version 1


Augustin Lefevre, Francis Bach, Cédric Févotte. Itakura-Saito nonnegative matrix factorization with group sparsity. 36th International Conference on Acoustics, Speech, and Signal Processing (ICASSP), May 2011, Prague, Czech Republic. ⟨hal-00567344⟩



Record views


Files downloads