Convex Color Image Segmentation with Optimal Transport Distances
Langue
en
Communication dans un congrès
Ce document a été publié dans
Proceedings of the fifth International Conference on Scale Space and Variational Methods in Computer Vision, Proceedings of the fifth International Conference on Scale Space and Variational Methods in Computer Vision, International Conference on Scale Space and Variational Methods in Computer Vision (SSVM'15), 2015-05-31, Lège Cap Ferret. 2015-05-31p. 256-269
Résumé en anglais
This work is about the use of regularized optimal-transport distances for convex, histogram-based image segmentation. In the considered framework, fixed exemplar histograms define a prior on the statistical features of the ...Lire la suite >
Mots clés en anglais
Optimal transport
Wasserstein distance
Sinkhorn distance
convex optimization
image segmentation