Color Image Segmentation Using a Model-Based Clustering and a MFA-EM Algorithm
In this paper we present a statistical model-based approach to the color image segmentation. A novel deterministic annealing EM and mean field theory are used to estimate the posterior probability of each pixel and the parameters of the Gaussian mixture model which represents the multi-colored objects statistically. Image segmentation is carried out by clustering each pixel into the most probable component Gaussian. The experimental results show that the mean field annealing EM provides a global optimal solution for the ML parameter estimation and the real images are segmented efficiently using the estimates computed by the maximum entropy principle and men field theory.
KeywordsImage Segmentation Gaussian Mixture Model Markov Random Field Block Image Maximum Entropy Principle
- M. Saeed, W. C. Karl, T. Q. Nguyen, and H. R. Rabiee, “A new multiresolution algorithm for image segmentation,” International Conference on ASSP, vol. 5, pp. 2753–2733, 1998.Google Scholar
- G. J. McLachlan, S.K. Nguyen, G.J. Galloway, and D. Wang, “Clustering of magnetic resonance images,” Technical Report, Department of Mathematics, University of Queensland, 1998.Google Scholar
- T. Hofmann and J. M. Buhman, “Pairwise data clustering by deterministic annealing,” IEEE Transactions on PAMI, vol. 19, no. 1, pp. 1–13, 1998.Google Scholar