Image Segmentation Method Combines MPM/MAP Algorithm and Geometric Division

Abstract

A novel image segmentation algorithm based on a Bayesian framework is studied in this paper. We presents a new region and statistics based approach, which combines Voronoi tessellation technique and Maximum a posterior / Maximization of the posterior marginal (MAP /MPM) algorithm. The image domain is partitioned into a group of sub-regions by Voronoi tessellation, each of which is a component of homogeneous regions. And the image is modeled on the supposition that the intensities of pixels in each homogenous region satisfy an identical and independent gamma distribution. The initial segmentation is applied to obtain number of the initial motions and the corresponding initial parameters of the image model. Then the parameters are updated by using the given parameter estimation method. A fast estimation procedure for the posterior marginals is added to the MAP algorithm. The experiment results show that the proposed algorithm here is effective.