Generalized Voronoi Tessellations for Vector-Valued Image Segmentation

We address the issue of low-level segmentation for vectorvalued images, focusing on color images. The proposed approach relies on the formulation of the problem as a generalized Voronoi tessellation of the image domain. In this context, the issue is transferred to the definition of an appropriated pseudo-metric and the selection of a set of sources. Two types of pseudo-metrics are considered; the first one is based on energy minimizing paths and the second is associated to the families of nested partitions of the image domain. We discuss specific applications of our approach to pre-segmentation, edge detection and hierarchical segmentation on color images.