Dictionary Learning for Sparse Representations: A Pareto Curve Root Finding Approach

Abstract : A new dictionary learning method for exact sparse representation is presented in this paper. As the dictionary learning methods often iteratively update the sparse coefficients and dictionary, when the approximation error is small or zero, algorithm convergence will be slow or non-existent. The proposed framework can be used in such a setting by gradually increasing the fidelity of the approximation. This technique has previously been used for the convex sparse representations. It has been extended here to the non-convex dictionary learning problem by allowing the dictionary be modified.
Type de document :
Chapitre d'ouvrage
Vigneron, Vincent and Zarzoso, Vicente and Moreau, Eric and Gribonval, Rémi and Vincent, Emmanuel. Latent Variable Analysis and Signal Separation, 6365, Springer Berlin / Heidelberg, pp.410-417, 2010
Liste complète des métadonnées

https://hal.inria.fr/inria-00567528
Contributeur : Jules Espiau de Lamaestre <>
Soumis le : lundi 21 février 2011 - 14:11:01
Dernière modification le : jeudi 26 octobre 2017 - 16:34:02

Identifiants

  • HAL Id : inria-00567528, version 1

Collections

Citation

Mehrdad Yaghoobi, Mike E. Davies. Dictionary Learning for Sparse Representations: A Pareto Curve Root Finding Approach. Vigneron, Vincent and Zarzoso, Vicente and Moreau, Eric and Gribonval, Rémi and Vincent, Emmanuel. Latent Variable Analysis and Signal Separation, 6365, Springer Berlin / Heidelberg, pp.410-417, 2010. 〈inria-00567528〉

Partager

Métriques

Consultations de la notice

47