基于顶点重要性骨架图的视点快速选择算法  

A skeleton based vertex significance map for rapid viewpoint selection

作　　者：金凯[1] 石振锋[2] 牛夏牧[3] 

机构地区：[1]哈尔滨工业大学建筑学院,哈尔滨150001 [2]哈尔滨工业大学数学系,哈尔滨150001 [3]哈尔滨工业大学计算机科学与技术学院,哈尔滨150001

出　　处：《哈尔滨工业大学学报》2012年第5期106-110,共5页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金资助项目(60832010;61100187);中国博士后研究基金资助项目(2011M500666);哈工大科研创新基金资助项目(HIT.NSRIF.2010046)

摘　　要：针对三维模型形状分析与处理、图形绘制、场景理解与优化布局等应用中的视点选择问题,利用三维模型线形骨架定义源模型顶点的重要性,给出了视点信息量的计算方法.采用Loop子分模板,对选定的初始视点集所形成的三角网格进行迭代式子分,在给定阈值的前提下,给出了三维模型最佳视点和最不利视点的快速选择.理论分析和实验结果表明:基于顶点重要性构建的骨架图,利用子分模板可快速地实现三维模型最佳视点和最不利视点的选择,算法执行效率高,能在最佳视点中提供更多的全局拓扑知觉信息.Aiming at the problem that viewpoint selection are widely used in many applications,such as,3D object visual shape analysis and processing,graphics drawing,scene rendering and optimal layout,we define a new skeleton map with the vertex importance based on curve-skeleton of 3D object and present a computation method for the information of viewpoint.Based on Loop subdivision stencil,the final best viewpoint and worst viewpoint can be rapidly achieved under a certain threshold using iterative subdivision for the triangular mesh generating from the selected initial viewpoints.Both analytical and experimental results show that the new skeleton map can effectively integrate into viewpoint selection and provide the functionality of rapid selection for the final best and worst viewpoints.The run time efficiency and maximum visual topological perception information is also demonstrated by these experiments and result analysis.

关 键 词：拓扑知觉 模型骨架 骨架图 顶点重要性 视点选择 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]