K-WEB: Nonnegative dictionary learning for sparse image representations

Abstract : This paper presents a new nonnegative dictionary learning method, to decompose an input data matrix into a dictionary of nonnegative atoms, and a representation matrix with a strict l0-sparsity constraint. This constraint makes each input vector representable by a limited combination of atoms. The proposed method consists of two steps which are alternatively iterated: a sparse coding and a dictionary update stage. As for the dictionary update, an original method is proposed, which we call K-WEB, as it involves the computation of k WEighted Barycenters. The so designed algorithm is shown to outperform other methods in the literature that address the same learning problem, in different applications, and both with synthetic and "real" data, i.e. coming from natural images.
Type de document :
Communication dans un congrès
IEEE International Conference on Image Processing (ICIP), Sep 2013, Melbourne, Australia. 2013
Liste complète des métadonnées

Littérature citée [9 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-00876018
Contributeur : Marco Bevilacqua <>
Soumis le : mercredi 23 octobre 2013 - 14:39:36
Dernière modification le : jeudi 15 novembre 2018 - 11:57:53
Document(s) archivé(s) le : vendredi 24 janvier 2014 - 04:25:48

Fichier

icip13_main.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00876018, version 1

Citation

Marco Bevilacqua, Aline Roumy, Christine Guillemot, Marie-Line Alberi Morel. K-WEB: Nonnegative dictionary learning for sparse image representations. IEEE International Conference on Image Processing (ICIP), Sep 2013, Melbourne, Australia. 2013. 〈hal-00876018〉

Partager

Métriques

Consultations de la notice

594

Téléchargements de fichiers

181