基于扩散张量成像的婴幼儿大脑图像自动分割

发布时间：2018-02-01 03:35

本文关键词： 婴儿儿童磁共振成像扩散张量成像大脑分割　出处：《中国医学影像技术》2017年08期 　论文类型：期刊论文

【摘要】：目的探讨基于扩散张量成像(DTI)的婴幼儿大脑图像自动分割方法的可行性及价值。方法提出一种基于DTI图像的婴幼儿大脑的分割方法。该方法主要分为2个阶段:①利用水的分布,提取脑脊液(CSF);②利用水在神经元中的各向异性扩散,提取白质(WM),继以区分灰质(GM)成分。结果通过本研究设计的特征选取方法可筛选出有效的DTI特征组合。第1步以平均扩散率(MD)和第3个特征值(L3)为组合特征提取CSF,第2步以各向异性分数(FA)和L3为组合特征提取WM和GM,可获得最高的平均相似性。经2步分割可成功进行婴幼儿大脑图像分割并获得满意的分割效果。结论基于DTI的婴幼儿大脑图像自动分割方法合理、可行,具有较高的分割精确度。
[Abstract]:Objective to study DTI based on diffusion Zhang Liang imaging. Methods A method based on DTI image was proposed to segment the brain of infants and young children. The method was divided into two stages: 1: 1 distribution of water. Extraction of CSF from cerebrospinal fluid; (2) extraction of white matter (WM) by anisotropic diffusion of water in neurons. Results the effective combination of DTI features can be selected by the method of feature selection designed in this study. The first step is based on the average diffusivity (MD) and the third eigenvalue (L3). Extract CSF for combined features. The second step is to extract WM and GM with anisotropic fractions (FA) and L3 as combined features. The highest average similarity can be obtained. After two steps of segmentation, the infant brain image can be segmented successfully and the result is satisfactory. Conclusion the automatic segmentation method of infant brain image based on DTI is reasonable and feasible. It has high segmentation accuracy.
【作者单位】：中南大学基础医学院生物医学工程系;邵阳学院临床医学系;
【基金】：湖南省教育厅科研项目基金(14C1042)
【分类号】：R445.2
【正文快照】： 精确、自动地大脑分割技术可应用于婴幼儿大脑DTI的各参数图,包括FA、平均扩散率(mean　diffus-损伤、脑功能重塑、大脑发育的相关认知科学研究。尽ity,MD)、扩散张量的3个特征值(L1、L2、L3,L1管常规MR图像在形态上的分辨率较高,但仍存在诸L2L3)及特征向量(V1、V2、V3)、原

【相似文献】