A generalized Swendsen-Wang algorithm for Bayesian nonparametric joint segmentation of multiple images

Jessica Sodjo (1) , Audrey Giremus (1) , Nicolas Dobigeon (2, 3) , Jean-François Giovannelli (1)

1 IMS - Laboratoire de l'intégration, du matériau au système
2 IRIT-SC - Signal et Communications
3 Toulouse INP - Institut National Polytechnique (Toulouse)

A generalized Swendsen-Wang (GSW) algorithm is proposed for the joint segmentation of a set of multiple images sharing, in part, an unknown number of common classes. The class labels are a priori mod-eled by a combination of the hierarchical Dirichlet process (HDP) and the Potts model. The HDP allows the number of regions in each image and classes to be automatically inferred while the Potts model ensures spatially consistent segmentations. Compared to a classical Gibbs sampler, the GSW ensures a better exploration of the posterior distribution of the labels. To avoid label switching issues, the best partition is estimated using the Dahl's criterion.

Mots clés: Index Terms— Image segmentation; Dirichlet Process; Bayesian nonparametrics; Image segmentation; Swendsen-Wang algorithm; Potts model

Domaines: Traitement du signal et de l'image [eess.SP]

Jessica Sodjo, Audrey Giremus, Nicolas Dobigeon, Jean-François Giovannelli. A generalized Swendsen-Wang algorithm for Bayesian nonparametric joint segmentation of multiple images. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Mar 2017, La Nouvelle Orléans, LA, United States. pp.1882-1886, ⟨10.1109/ICASSP.2017.7952483⟩. ⟨hal-01695104⟩