采用一种新的马尔可夫模型进行癌症医学图像分割

Using A New Markov Model in Carcinoma Image Segmentation

将数据驱动的蒙特卡罗型马尔可夫链(DDMCMC)这一新的马尔可夫模型运用于分割癌症医学图像.这一图像分割过程为:先将需要处理的空间分割成3层;创建平稳的马尔可夫链;采用数据驱动的方法运算得到启发式的信息;运用适合的数学原理来选取图像粒子.通过对比可知采用DDMCMC方法比采用其它算法在癌症医学图像分割方面更具有优点.

In this paper, a new Markov model named Data Driven Markov Chain Monte Carlo (DDMCMC) is developed to segment carcinoma images. The segmentation process will be divided into four steps. First, this method divides the solution space into three layers. Second, the balanced Markov chains are created. Then data driven techniques are used to compute the heuristic information. Finally, a math principle is proposed to choose scene particles. Through the comparison between DDMCMC and other arithmetic, this method shows its advantages in carcinoma image segmentation.