摘要
本文利用脑图谱的先验知识并结合水平集等算法实现对脑MR图像的初步分割。主要步骤:(1)选取数字脑图谱,对图谱进行预处理;(2)实现图谱与脑MR图像的配准;(3)利用图谱提供的轮廓信息对水平集算法进行初始化,完成颅骨和脑脊液的提取以及脑白质和脑灰质的分割。实验结果表明,利用脑图谱提供的信息可有效解决水平集算法初始化问题,缩小求解空间,减少迭代次数,该方法具有较好的鲁棒性。
In this paper, a MR image segmentation method was presented by unifying prior knowledge of atlas and level set arithmetic. The main processes： （1） choose a digital brain atlas and preprocess it; （2） perform the atlas and the MR image registration; （3） initialize level set arithmetic based an contour information of the atlas. Realize skull and cerebraspinal （CSF） extraction and achieve gray mttter （GM） and white matter （WM） segmentation. Experiment results showed that the problem of level set initializations was solved based on the information of alias and it could reduce the solution space and iteration. This method is characterized by robustness.
出处
《中国医学物理学杂志》
CSCD
2006年第5期329-332,344,共5页
Chinese Journal of Medical Physics
关键词
脑图谱
水平集
MR图像分割
brain atlas
level set
MR image segmentation
