医学体数据中四面体化方法的研究进展

Research progress in tetrahedral mesh generation from medical volume data

四面体化方法一直是网格生成研究的热点,然而将其应用于医学体数据的器官几何建模仍存在诸多难点。根据医学体数据的结构特点,首先阐述了Marching cubes重构器官表面的原理和研究新进展,然后以有限元方法为应用背景,按照体数据的两种处理方式,从基于表面建模和基于体素建模两方面进行讨论,分别研究与比较三种四面体化方法在不同输入情况下构建器官几何模型的特点和适用范围,通过改进与实现几个不同种类的算法,分析各自的优缺点以及仍然存在的问题,探讨相应的改进措施。其中重点归纳了基于Delaunay准则的不同算法构建器官四面体网格模型的新思路以及在保持边界和Sliver去除方面的改进措施。最后探讨了四面体化方法在该类几何建模中的研究热点和研究方向。

Tetrahedral mesh generation has always been a hot issue in the area of mesh generation,but its application in geometrical construction of organs from medical volume data is still very difficult.This paper first gave an introduction to the state-of-art of the Marching cubes which was the most popular isofacing algorithm.Then,according to different preprocessing for medical volume data,summarized and compared three types of tetrahedral mesh generation in their developments,extension and limitations（including the attempts to resolve its limitations） through classic examples.Especially,analyzed and reviewed some novel algorithms based on Delaunay in detail.Along existing methods,predicted some potential directions for future research finally.