Joint segmentation of multiple images with shared classes: a Bayesian nonparametrics approach

Abstract : A combination of the hierarchical Dirichlet process (HDP) and the Potts model is proposed for the joint segmenta-tion/classification of a set of images with shared classes. Images are first divided into homogeneous regions that are assumed to belong to the same class when sharing common characteristics. Simultaneously, the Potts model favors configurations defined by neighboring pixels belonging to the same class. This HDP-Potts model is elected as a prior for the images, which allows the best number of classes to be selected automatically. A Gibbs sampler is then designed to approximate the Bayesian estimators, under a maximum a posteriori (MAP) paradigm. Preliminary experimental results are finally reported using a set of synthetic images.
Type de document :
Communication dans un congrès
IEEE Statistical Signal Processing Workshop (SSP), Jun 2016, Palma de Majorque, Spain. 2016 IEEE Statistical Signal Processing Workshop (SSP), 〈10.1109/SSP.2016.7551735〉
Liste complète des métadonnées

Littérature citée [13 références]  Voir  Masquer  Télécharger

https://hal.archives-ouvertes.fr/hal-01695100
Contributeur : Jessica Sodjo <>
Soumis le : lundi 29 janvier 2018 - 08:04:38
Dernière modification le : mercredi 12 septembre 2018 - 17:46:03
Document(s) archivé(s) le : vendredi 25 mai 2018 - 13:37:42

Fichier

07551735.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Relations

Citation

Jessica Sodjo, Audrey Giremus, Francois Caron, Jean-François Giovannelli, Nicolas Dobigeon. Joint segmentation of multiple images with shared classes: a Bayesian nonparametrics approach. IEEE Statistical Signal Processing Workshop (SSP), Jun 2016, Palma de Majorque, Spain. 2016 IEEE Statistical Signal Processing Workshop (SSP), 〈10.1109/SSP.2016.7551735〉. 〈hal-01695100〉

Partager

Métriques

Consultations de la notice

125

Téléchargements de fichiers

31