A novel image segmentation method based on improved MRF model

This paper proposes a novel segmentation method based on improved Markov Random Field(MRF) model, which integrates priori and boundary information of the image. First, proposes a novel prior energy function which uses pixel intensity and boundary information of the image simultaneously for the first time, it can give higher segmentation accuracy while maintaining a good boundary. Then, introduces a novel energy minimization method namely Simulated Annealing With Probability Table (SAP) into the MRF model for the first time, it can greatly enhance the speed of global optimization while obtaining the segmentation accuracy. Experiments on simulated image and real clinical images show that this model is robust, accurate and efficient, especially for the weak boundary and concave region.