3-D Voronoi diagram: application to segmentation

Computational geometry provides many solutions to imaging problems, mainly for 3-D image compression, segmentation, and measurement. The authors present a method to partition a volume data with Voronoi polyhedra structured in a graph environment. A dynamic construction of the 3-D Voronoi diagram is defined using an interactive method, guided by the image information. The developed process has been applied to confocal laser scanning microscopy. The authors developed also two other geometrical models: the first one is the octree model and the second one is the Delaunay model, dual of the Voronoi diagram.< >