A new unsupervised image segmentation algorithm based on deterministic annealing EM

Author

Zhong, Jiaqiang ; Wang, Runsheng

Author_Institution

ATR Nat. Labs, Nat. Univ. of Defense Technol., Changsha, China

Volume

1

fYear

2003

Firstpage

600

Abstract

A new unsupervised image segmentation algorithm based on deterministic annealing EM (DAEM) is proposed in this paper. The method is based on maximum likelihood (ML) estimation. Image is considered as a mixture of multi-variant normal densities and the number of densities is assumed to know. In order to obtain the parameters of densities, deterministic annealing EM algorithm is introduced. In DAEM algorithm, the EM process is reformulated as the problem of minimizing the thermodynamic free energy by using a statistical mechanics analogy. Thus, The DAEM algorithm can overcome the local maximize problem of general EM algorithm. The proposed method is successfully applied to image segmentation experiments.

Keywords

deterministic algorithms; image segmentation; maximum likelihood estimation; minimisation; simulated annealing; statistical mechanics; deterministic annealing EM algorithm; image segmentation algorithm; maximum likelihood estimation; multivariant normal densities; statistical mechanics analogy; thermodynamic free energy minimization; Annealing; Convergence; Image processing; Image segmentation; Maximum likelihood estimation; Parameter estimation; Partitioning algorithms; Robots; Signal processing algorithms; Thermodynamics;

fLanguage

English

Publisher

ieee

Conference_Titel

Robotics, Intelligent Systems and Signal Processing, 2003. Proceedings. 2003 IEEE International Conference on

Print_ISBN

0-7803-7925-X

Type

conf

DOI

10.1109/RISSP.2003.1285642

Filename

1285642