A hybrid rotational matrix extraction method for soft tissue deformation modeling
A hybrid rotational matrix extraction method for soft tissue deformation modeling
复制标题
DOI：
10.1109/icma.2015.7237575
复制
发表时间：
2015-08
期刊：
影响因子：
--
通讯作者：
Xiufen Ye;Jianguo Zhang;Peng Li
中科院分区：
文献类型：
--
作者：
Xiufen Ye;Jianguo Zhang;Peng Li
本文主要研究了共旋有限元法中旋转矩阵的提取方法。该矩阵用于在软体变形模拟过程中的每个时间步进行刚度翘曲装配。QR分解和极坐标分解是计算机图形学中常用的两种分解方法。前者精度相对较高，但耗时较长，后者精度较低，但效率较高，本文对它们的实时性进行了分析。我们的混合旋转矩阵提取方法具有高效的数据结构，它结合了上述两种方法，使变形结果和实时性能之间的平衡。此外，还考虑了虚拟手术系统的具体情况，其中变形通常发生在局部接触区域，作为加速因子。实验结果表明，混合旋转矩阵提取方法可以保证变形模拟结果的准确性和快速的计算效率。
In this paper, we mainly investigated the methods of extracting rotational matrix in co-rotational finite element method (FEM). The matrix is used to perform stiffness warping assembly at every time-step in a soft body deformation simulation procedure. QR decomposition and polar decomposition are two well-known decomposition methods in computer graphics. The former is relative precise but time-consuming, while the latter is less precise but more efficient, and their real-time performances were analyzed in this paper. Our hybrid rotational matrix extraction method with efficient data structure, which combines the two methods mentioned above to make a balance between deformation results and real-time performance. In addition, the specific situation of a virtual surgery system has also been taken into account, in which the deformation usually occurs in a local contacted area, acts as an acceleration factor. The experimental results show that the hybrid rotational matrix extraction method can guarantee both accurate deformation simulation results and fast computation efficiency.