A Novel Segmentation Method for Left Ventricular from Cardiac MR Images Based on Improved Markov Random Field Model

In this paper, we propose a improved Markov Ran- dom Field (MRF) segmentation model, which integrates region, priori knowledge and boundary information of the image, for segmenting left ventricle (LV) boundary from cardiac MR image. The proposed model incorporates geometry shape boundary information, and improves the objective function of traditional MRF model. Furthermore, Chaotic Simulated Annealing (CSA) algorithm is introduced to solve the MRF model for the first time. Since CSA algorithm introduces chaos ergodicity mechanism, it can take advantage of Chaos Algorithm (COA) and Simulated Annealing (SA) algorithm in the search process. CSA algorithm can not only avoid the limitations of mathematical optimization methods, but also greatly enhance the speed of global optimiza- tion. Experiments on clinical cardiac MR images show that the improved MRF model has high performance on segmenting LV boundary. The evaluation results illustrate that this model is robust, accurate and efficient, especially for the weak boundary and concave region . contrast, the segmentation results of MRF model are not very satisfactory. In this paper, we propose a improved Markov Random Field (MRF) segmentation model, which integrates region, priori knowledge and boundary information of the image, for segmenting LV boundary from cardiac MR image. Often, the segmentation problem is cast using maximum a posteriori (MAP) criterion and the MAP estimates are obtained by SA algorithm. SA is a global optimization algorithm, and it can guarantee to find the optimal solution. But SA has some problems such as time consuming and slow convergence. In order to obtain the optimal solution, and improve the convergence rate, this paper first introduces CSA algorithm to MRF model. Since CSA algorithm introduces chaos ergodicity mechanism, it can take advantage of the global ergodicity of COA and heuristic rules of SA algorithm in the search process. CSA algorithm can not only avoid the limitations of mathe- matical optimization methods, but also greatly enhance the speed of global optimization. Finally, the proposed improved MRF model is applied to segment LV boundary from clinical cardiac MR image. Experiments show that the improved MRF model has high performance on segmenting LV boundary. The evaluation results illustrate that this model is robust, accurate and efficient, especially for the weak boundary and concave region.

[1]  P. Nanda,et al.  Constrained Markov Random Field Model for Color and Texture Image Segmentation , 2008, 2008 International Conference on Signal Processing, Communications and Networking.

[2]  Chunming Li,et al.  Distance Regularized Level Set Evolution and Its Application to Image Segmentation , 2010, IEEE Transactions on Image Processing.

[3]  Jayaram K. Udupa,et al.  A framework for evaluating image segmentation algorithms , 2006, Comput. Medical Imaging Graph..

[4]  Fang Liu,et al.  A New MRF Framework with Dual Adaptive Contexts for Image Segmentation , 2007, 2007 International Conference on Computational Intelligence and Security (CIS 2007).

[5]  Shi Zhang,et al.  A Knowledge-based Segmentation Method Integrating both Region and Boundary Information of Medical Images , 2008, 2008 International Conference on BioMedical Engineering and Informatics.

[6]  Tamás Szirányi,et al.  Picture segmentation with introducing an anisotropic preliminary step to an MRF model with cellular neural networks , 1996, Proceedings of 13th International Conference on Pattern Recognition.

[7]  Jue Wu,et al.  A Segmentation Model Using Compound Markov Random Fields Based on a Boundary Model , 2007, IEEE Transactions on Image Processing.

[8]  Bing Li,et al.  Active Contour External Force Using Vector Field Convolution for Image Segmentation , 2007, IEEE Transactions on Image Processing.

[9]  A. Fenster,et al.  Evaluation of Segmentation algorithms for Medical Imaging , 2005, 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference.

[10]  Qianjin Feng,et al.  Cardiac MR Image Segmentation with Modified Active Contour Model , 2007, 2007 IEEE/ICME International Conference on Complex Medical Engineering.

[11]  Hong Sun,et al.  An Unsupervised Segmentation Method Using Markov Random Field on Region Adjacency Graph for SAR Images , 2006, 2006 CIE International Conference on Radar.

[12]  John W. Woods,et al.  Compound Gauss-Markov random fields for image estimation , 1991, IEEE Trans. Signal Process..

[13]  P. K. Nanda,et al.  Color image segmentation using constrained compound Markov Random Field model and homotopy continuation method , 2008, 2008 First International Conference on Distributed Framework and Applications.

[14]  Lei Guo,et al.  A Novel MRF-Based Image Segmentation Algorithm , 2006, 2006 9th International Conference on Control, Automation, Robotics and Vision.

[15]  Ping Zhong,et al.  A New MRF Framework with Dual Adaptive Contexts for Image Segmentation , 2007 .