基于模糊水平集的脑肿瘤MR图像分割方法

Brain tumor MR image segmentation method based on fuzzy level set

针对传统水平集(Level Set)方法对脑肿瘤MR图像进行分割时易在弱边缘处产生泄露的问题,提出一种新的基于模糊水平集的脑肿瘤MR图像分割方法。采用模糊聚类算法对图像进行预分割,得到脑肿瘤MR图像的感兴趣区域;将聚类分割结果作为水平集演化的初始轮廓;利用聚类结果计算水平集演化的初始化条件和控制参数。算法执行效率得到了提高,并且克服了水平集演化依赖于初始化条件和控制参数且需要较多人工干预的缺陷,增加了方法的鲁棒性。实验结果表明,该方法鲁棒性强,能够快速、准确地分割出MR图像中的脑肿瘤,具有重要的临床意义。

Since the traditional level set method used to segment the brain tumor MR image is liable to generate the leakageat weak edges,a new brain tumor MR image segmentation method based on fuzzy level set is proposed,in which the fuzzy clus?tering algorithm is used to pre?segment the image to get the interest area of brain tumor MR image,the result of clustering seg?mentation is taken as the initial contour of level set evolution,and then the clustering result is used to calculate the initial condi?tions and control parameters of the level set evolution. This method can improve the algorithm execution efficiency;overcomethe defects that the level set evolution relies on initial conditions,control parameters and more manual intervention;and im?prove the robustness. The experimental results show this method has high robustness,can segment the brain tumor in MR imagequickly and accurately,and has important clinical significance.