Markov random field method is a very active research field in image segmentation. This paper introduces the relationship between a general theory based on Markov random field method and the images
and provides a general framework in image segmentation
including the construction of spatial and wavelet domain image models
the selection of the optimization criterion
calculation of the number of labeling
parameter estimation of image models and the realization of image segmentation. The applications of image segmentation are reviewed. And a few possible trends are discussed.