Stochastic Watershed Hierarchies
Résumé
We present a segmentation strategy which first constructs a hierarchy, i.e. a series of nested partitions. A coarser partition is obtained by merging adjacent regions in a finer partition. The strength of a contour is then measured by the level of the hierarchy for which its two adjacent regions merge. Various strategies are presented for constructing hierarchies which highlight specific features of the image. The core of the paper deals with the stochastic watershed hierarchy. It establishes the probabiity of a contour to appear if a number of random germs are introduced in an image. According to the probability law governing the placement and the shape of the germs, very expressive hierarchies are produced. The last part shows how the hierarchies lead to a final segmentation.
Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...