基于形状直径函数的三维模型集一致性分割算法研究  

Consistency Segmentation Algorithm of 3D Model Set Based on Shape Diameter Function

作　　者：贾晖 张建刚[2] JIA Hui;ZHANG Jian-gang(School of Computer Science,Xi'an University of Posts and Telecommunications,Xi'an 710121,China;Technological Department of Power Station Information and Monitoring,Xi'an Thermal Power Research Institute Co.,Ltd.,Xi'an 710032,China)

机构地区：[1]西安邮电大学计算机学院,西安710121 [2]西安热工研究院有限公司电站信息及监控技术部,西安710032

出　　处：《价值工程》2019年第17期277-279,共3页Value Engineering

基　　金：陕西省教育厅专项科研计划项目“无监督的三维模型集有意义聚类分割技术研究”,项目编号:15JK1656;国家自然科学基金面上项目“面向图文混合的网络舆情新事件发现及其关联挖掘”,项目编号:61572399

摘　　要：为了有效提高对三维模型集进行一致性分割算法的准确率和效率,提出一种三维模型集一致性分割算法。该算法基于形状直径函数特征和谱聚类方法。首先计算模型各个面片的形状直径函数特征(SDF)特征;其次计算模型各面片之间的形状相似性,形成相似性矩阵。最后采用谱聚类算法对模型集进行一致性分割。将本文算法与相关方法进行对比,实验结果表明,本文算法能准确的对一致性分割数据集中的若干模型进行有意义的分割,使得分割部位具有一定的对应关系,且面片平均划分准确率较好。In order to improve the accuracy and efficiency of the algorithm of 3D model set consistency segmentation, an algorithm of 3D model set consistency segmentation is proposed. This algorithm is based on Shape diameter function(SDF) and spectral clustering. At first, The face SDF characteristics of each models in the model set is extracted;and then The similarity between each face of the model is calculated, and the similarity matrix is formed. At last utilize the spectral clustering algorithm to segment the model set consistently. The algorithm in this paper is compared with the relevant methods, the experimental results show that the proposed algorithm can accurately segment several models in a consistent data set, make the segmentation parts have a certain correspondence, and the average segmentation accuracy is better.

关 键 词：三维模型 一致性分割 形状直径函数 谱聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]