3d morphological segmentation for image sequence processing

The aim of this paper is to present a method for segmenting image sequences. This method is based on a three dimensional (3D) morphological segmentation. As morphological tools are very efficient in order to obtain a segmentation based on the real objects of the scene, the proposed scheme extends to three dimensions a morphological segmentation method for still images. This extension arises different problems due to the fact that we are not dealing with three dimensional objects but with a sequence of images taken at discrete instants on the time dimension. Two different solutions are proposed to deal with the 3D morphological segmentation: interpolation of the lattice of the image sequence and a modification of the basic morphological tool for segmentation: the watershed algorithm.