Markov random field model and expectation of maximization for images segmentation

Image segmentation is a significant issue in image processing. Among the various models and approaches that have been developed, some are commonly used the Markov random field (MRF) model, statistical techniques MRF. In this study a Markov random field proposed is based on an expectation-maximization (EM) modified (EMM) model. In this paper, the local optimization is based on a modified EM method for parameter estimation and the iterative conditional model (ICM) method for finding the solution given a fixed set of these parameters. To select the combination strategy, it is necessary to carry out a comparative study to find the best result. The effectiveness of our proposed methods has been proven by experimentation. We have applied this segmented algorithm to different types of images, exhibiting the algorithm's image segmentation strength with its best values criteria for EM statics and other methods.