• français
    • English
  • français 
    • français
    • English
  • Connexion
JavaScript is disabled for your browser. Some features of this site may not work without it.
Accueil

Afficher

Cette collectionPar Date de CréationAuteursTitresSujetsNoms de revueToute la baseCentres de recherche & CollectionsPar Date de CréationAuteursTitresSujetsNoms de revue

Mon compte

Connexion

Statistiques

Afficher les statistiques d'usage

Wasserstein Loss for Image Synthesis and Restoration

Thumbnail
Ouvrir
WassersteinFidelity.pdf (5.174Mb)
Date
2016
Indexation documentaire
Traitement du signal
Subject
Super-resolution; Total variation; Denoising; Generalized Gaussian Distributions; Optimal transport; Wasserstein; impulse noise removal
Nom de la revue
SIAM Journal on Imaging Sciences
Volume
9
Numéro
4
Date de publication
2016
Pages article
1726-1755
Nom de l'éditeur
Society for Industrial and Applied Mathematics
DOI
http://dx.doi.org/10.1137/16M1067494
URI
https://basepub.dauphine.fr/handle/123456789/17015
Collections
  • CEREMADE : Publications
Métadonnées
Afficher la notice complète
Auteur
Tartavel, Guillaume
Peyré, Gabriel
Gousseau, Yann
Type
Article accepté pour publication ou publié
Résumé en anglais
This paper presents a novel variational approach to impose statistical constraints to the output of both image generation (to perform typically texture synthesis) and image restoration (for instance to achieve denoising and super-resolution) methods. The empirical distributions of linear or non-linear descriptors are imposed to be close to some input distributions by minimizing a Wasserstein loss, i.e. the optimal transport distance between the distributions. We advocate the use of a Wasserstein distance because it is robust when using discrete distributions without the need to resort to kernel estimators. We showcase different estimators to tackle various image processing applications. These estimators include linear wavelet-based filtering to account for simple textures, non-linear sparse coding coefficients for more complicated patterns, and the image gradient to restore sharper contents. For applications to texture synthesis, the input distributions are the empirical distributions computed from an exemplar image. For image denoising and super-resolution, the estimation process is more difficult; we propose to make use of parametric models and we show results using Generalized Gaussian Distributions.

  • Accueil Bibliothèque
  • Site de l'Université Paris-Dauphine
  • Contact
SCD Paris Dauphine - Place du Maréchal de Lattre de Tassigny 75775 Paris Cedex 16

 Cette création est mise à disposition sous un contrat Creative Commons.