Learning Dictionaries as Sums of Kronecker Products

Abstract : The choice of an appropriate dictionary is a crucial step in the sparse representation of a given class of signals. Traditional dictionary learning techniques generally lead to unstructured dictionaries which are costly to deploy and do not scale well to higher dimensional signals. In order to overcome such limitation, we propose a learning algorithm that constrains the dictionary to be a sum of Kronecker products of smaller sub-dictionaries. A special case of the proposed structure is the widespread separable dictionary. This approach, named SuKro, is evaluated experimentally on an image denoising application.
Type de document :
Document associé à des manifestations scientifiques
SPARS 2017 - Signal Processing with Adaptive Sparse Structured Representations workshop, Jun 2017, Lisbon, Portugal. 〈http://spars2017.lx.it.pt/〉
Liste complète des métadonnées

Littérature citée [17 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01514044
Contributeur : Cassio Dantas <>
Soumis le : mardi 25 avril 2017 - 15:43:19
Dernière modification le : mercredi 16 mai 2018 - 11:24:14
Document(s) archivé(s) le : mercredi 26 juillet 2017 - 14:26:45

Fichier

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

Identifiants

  • HAL Id : hal-01514044, version 1

Citation

Cassio Fraga Dantas, Rémi Gribonval, Renato Lopes, Michele Da Costa. Learning Dictionaries as Sums of Kronecker Products. SPARS 2017 - Signal Processing with Adaptive Sparse Structured Representations workshop, Jun 2017, Lisbon, Portugal. 〈http://spars2017.lx.it.pt/〉. 〈hal-01514044〉

Partager

Métriques

Consultations de la notice

689

Téléchargements de fichiers

156