摘要
人脑形态研究在许多领域都有日益广泛的应用。将人的大脑从 MR图像体数据中提取出来是脑的可视化、形态学分析、脑图象配准等多种工作的重要基础。本文结合解剖学知识、用基于灰度梯度的极坐标搜索方法从人脑 MR图像自动剔除大脑皮层和颅骨及其它非脑成分 ,仅保留大脑部分。此后 ,对于 3D图像遍历 ,我们使用移动立方体法根据全部 2 56种构型 ,通过查表法把位于大脑表面上的像素与大脑内部及图像背景分开 ,得到准确的大脑表面轮廓。然后计算该表面轮廓的各个小三角形面片法线 ,结合适当的光照模型实现大脑表面形态的真实再现。大脑的提取工作是在轴向的方向完成的。但与轴向图正交的冠状图和矢状图显示结果表明 。
The study of morphology of humn brain has increasing applications in many areas. Extraction of cerebrum from raw MRI volume data is an important work for visualization of human brain, analysis of morphology and image registration. Using priors of anatomy, agrey level gradient based searching method in a polar coordinate system to remove cortex, skull and other no cerebrum structures automatically from MR image of human brain is described in this paper. Then, For 3D image roaming, a mothod of cubic displacement was used based on all 256 possible configurations, a look up table is used to get the accurate contour of brain surface by distinguishing pixels of brain surface from pixels of internal brain and the background. Then, normals of all the triangle facets of the surface contour are calculated, to re display the surface morphology, combining with appropriate illumination model. Although the extraction of cerebrum is implemented in transverse slices, the reconstructed coronal and sagittal slices show the work is very succesful.\;
出处
《北京生物医学工程》
2000年第4期200-205,共6页
Beijing Biomedical Engineering
基金
北京市教委科技发展计划项目
关键词
NMR
成像
表面重建
大脑
表面形态
新方法
MR image, Removal of cortex and skull, Surface rendering, Morphology, Contour
