基于C-V模型和MRF的脑部MRI分割

Brain MRI Segmentation Based on C-V Model and MRF

为准确分割脑部磁共振图像(MRI)的灰质、白质和背景,提出一种基于C-V模型和马尔可夫随机场的全自动分割方法。采用C-V模型与形态学相结合的方法对脑MRI进行预处理,去除多余脑组织,获得待分割图像。引入灰度场局部熵的思想对惩罚因子进行估计,利用马尔可夫随机场模型建模实现脑灰白质的分割,并运用形态学方法获得最终分割结果。对96幅IBSR图像和46幅临床图像进行实验,结果表明,该方法能够实现脑部MRI灰白质的全自动分割,且具有较好的分割精度和较快的处理速度。

In order to get gray matter, white matter and background from brain Magnetic Resonance lmage（MRl） accurately, an automatic segmentation method based on C-V model and Markov Random Field（MRF） is proposed. It uses C-V algorithm and morphology to preprocess the original image and remove the unnecessary brain tissue, and the image to be segmented is got. It introduces the local entropy of the gray scale field to estimate penalty factor and Markov random field model is used to achieve the segmentation of gray matter and white matter. The segmentation result is obtained by morphological methods. Experiments are carried out on 96 pieces of IBSR images and 46 pieces of clinical images using this method, results show that the proposed method can achieve the automatic segmentation of the brain MRI and has higher accuracy as well as faster processing speed than before.