Image Segmentation Based On Energy Optimization Of Probabilistic Graphical Models

Image segmentation plays an important role in image processing.Image is a variation complex system,thus,the optimal segmentation of the system is unknown for the computer,and it is difficult to obtain the optimal solution with an explicit mathematical model.In particular,it is more difficult to get the optimal state solution for complex images.Therefore,the researchers proposed many image segmentation algorithms so as to segment the target images quickly and accurately.However,there are still some shortcomings for these segmentation algorithms while the target images have been segmented.So there is no perfect solution to image segmentation.An image segmentation algorithm of energy minimization based on graphical model is presented in this paper.The roughness correction algorithm based on Markov Random Field,level set algorithm and Grab Cut method are applied to this algorithm.The image can be considered as the undirceted probabilistic graphical model,i.e.,Markov Random Field model.The connection between the image pixels is mapped as the connection between the edge and node in Markov Random Field by graphical models,and energy function is constructed according to the priori knowledge of the image by Energy Minimization model.Thus,the problem of image segmentation can be converted to the problem of solving the minimum value of energy function.In this paper,the minimum value of energy function is solved by Grab Cut method,then the target image is segmented.The effect of image segmentation is not affected by image size in the proposed image segmentation algorithm in this paper,and the stability and the velocity of image segmentation are improved.The effect of image segmentation has nothing to do with the initial contour of image in the proposed image segmentation algorithm in this paper,and the accuracy of segmentation and the running velocity are greatly improved.Therefore,the effect of image segmentation has been improved,and there are some advantages of accuracy and the running velocity for the proposed image segmentation algorithm in this paper.It may portend a bright application prospect for the proposedmethod of imagesegmentation in this paper.