Dictionary learning methods for single-channel source separation

Augustin Lefèvre 1
1 SIERRA - Statistical Machine Learning and Parsimony
DI-ENS - Département d'informatique de l'École normale supérieure, ENS Paris - École normale supérieure - Paris, Inria Paris-Rocquencourt, CNRS - Centre National de la Recherche Scientifique : UMR8548
Abstract : In this thesis we provide three main contributions to blind source separation methods based on NMF. Our first contribution is a group-sparsity inducing penalty specifically tailored for Itakura-Saito NMF. In many music tracks, there are whole intervals where only one source is active at the same time. The group-sparsity penalty we propose allows to blindly indentify these intervals and learn source specific dictionaries. As a consequence, those learned dictionaries can be used to do source separation in other parts of the track were several sources are active. These two tasks of identification and separation are performed simultaneously in one run of group-sparsity Itakura-Saito NMF. Our second contribution is an online algorithm for Itakura-Saito NMF that allows to learn dictionaries on very large audio tracks. Indeed, the memory complexity of a batch implementation NMF grows linearly with the length of the recordings and becomes prohibitive for signals longer than an hour. In contrast, our online algorithm is able to learn NMF on arbitrarily long signals with limited memory usage. Our third contribution deals user informed NMF. In short mixed signals, blind learning becomes very hard and sparsity do not retrieve interpretable dictionaries. Our contribution is very similar in spirit to inpainting. It relies on the empirical fact that, when observing the spectrogram of a mixture signal, an overwhelming proportion of it consists in regions where only one source is active. We describe an extension of NMF to take into account time-frequency localized information on the absence/presence of each source. We also investigate inferring this information with tools from machine learning.
Document type :
Complete list of metadatas

Cited literature [121 references]  Display  Hide  Download

Contributor : Abes Star <>
Submitted on : Tuesday, February 20, 2018 - 11:53:05 AM
Last modification on : Wednesday, January 30, 2019 - 11:08:31 AM
Long-term archiving on : Monday, May 21, 2018 - 12:42:33 PM


Files produced by the author(s)


  • HAL Id : tel-00764546, version 2


Augustin Lefèvre. Dictionary learning methods for single-channel source separation. General Mathematics [math.GM]. École normale supérieure de Cachan - ENS Cachan, 2012. English. ⟨NNT : 2012DENS0051⟩. ⟨tel-00764546v2⟩



Record views


Files downloads