Statistical and spatial information association for clusters number and mean values estimation in image segmentation

A straightforward way to assess the number of clusters as well as their mean values for clustering-based gray scale image segmentation initialization purpose, is presented in this paper. It is based on an association of statistical and spatial information of the image so that, the final classes in the segmented image correspond to “meaningful” modes in an histogram built from the image characteristics. The experiments testify the efficiency of this method in case of synthetic and real images.