一种通用的局部内蕴对称检测算法  

A Generalized Algorithm for Partial Intrinsic Symmetry Detection

作　　者：姜巍[1] 徐凯[1] 程志全[1] 党岗[1] 

机构地区：[1]国防科学技术大学计算机学院,长沙410073

出　　处：《计算机辅助设计与图形学学报》2013年第7期974-979,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61272334;61103084;61202333)

摘　　要：局部内蕴对称在三维几何体中更具一般性,其检测对形状分析具有重要意义,但由于其求解的复杂性,使得局部内蕴对称的检测非常困难.文中面向三维网格模型,提出一种包括反射和平移等变换的通用局部内蕴对称检测算法.该算法使用选举和谱分析方法以及迭代精化对称对应矩阵,求解出更加精确、通用的局部内蕴对称性;进一步,根据对称检测结果对模型进行分割,得到有意义的分割结果.实验结果表明,采用文中算法可以正确地求解出通用的局部内蕴对称性,并能在通用对称属性引导下生成有意义的模型分割结果.As partial intrinsic symmetry is more general in the 3D shapes, this type symmetry detection is very important. However, it is difficult to detect the partial intrinsic symmetry properties, due to the intrinsic complexity of the solution. On the 3D meshes, this paper proposed one partial intrinsic symmetry detection algorithm in the generalized way, including the reflection and translation transformations. The algorithm employed the voting and spectral analysis technique. The Symmetry Correspondence Matrix was iteratively refined to compute more generalized partial intrinsic symmetries. Then, the symmetry detection results were applied at the meaningful segmentation. The experiments on meshes show that the algorithm could accurately compute the generalized partial intrinsic symmetries, and produce meaningful segmentation results guided by the detected symmetry properties.

关 键 词：局部内蕴对称 谱分析 选举方法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]