中文摘要：

本文针对自旋图计算效率随着三维模型顶点规模增大而降低的问题,基于串行的自旋图算法,给出了多线程的三维模型自旋图并行计算方法.本文首先给出了三维模型顶点上自旋图的定义及串行算法,然后详细描述了多线程的顶点自旋图并行计算方法,最后在实验结果部分分析对比了本文并行算法与串行算法的效率差异,以及三维模型的顶点规模、线程数目等因素对并行算法的运行时间、加速比、可扩放性等特性的影响.实验结果表明,与串行方法相比,本文提出的多线程并行算法具有显著的优势.

英文摘要：

We propose a novel multithreaded parallel algorithm to address efficiency problem of spin image, when the vertex scale of 3D models increases. Based on the serial algorithm,we firstly give the definition of spin image and its serial algorithm on the vertex of 3D models, and then give a detailed description of our multithreaded parallel algorithm. And furthermore, in the section of experiments, we compare the results, efficiency of serial algorithm and parallel algorithm, with the influence analysis of the thread number, the vertex number of 3D models on the run time,the speedup and the scalability. The results showed that,compared with the serial algorithm, our multi-threaded parallel algorithms proposed in this paper have a significant advantage.