基于决策图的三维模型无监督聚类算法
PDF下载 (17639)徐 欣 1,舒振宇 2,陈双敏 1*,辛士庆 3,屠长河 3.基于决策图的三维模型无监督聚类算法[J].宁波大学学报(理工版),2018,31(4):46-51.DOI:
XU Xin 1,SHU Zhen-yu 2,CHEN Shuang-min 1*,XIN Shi-qing 3,TU Chang-he 3.Unsupervised clustering algorithm on 3D model library via decision graph[J].Journal of Ningbo University(Natural Science & Engineering Edition),2018,31(4):46-51.DOI:
| Title: | Unsupervised clustering algorithm on 3D model library via decision graph |
| 作者: | 徐 欣 1, 舒振宇 2, 陈双敏 1*, 辛士庆 3, 屠长河 3 |
| Author(s): | XU Xin 1, SHU Zhen-yu 2, CHEN Shuang-min 1*, XIN Shi-qing 3, TU Chang-he 3 |
| 关键词: | 核函数; 无监督聚类; 三维模型; 决策图; 多特征融合 |
| Keywords: | kernel function; unsupervised clustering; 3D models; decision graph; fusing multiple features |
| 分类号: | TP301.6 |
| 文献标识码: | A |
| 摘要: | 针对三维模型的无监督聚类问题, 目前广泛采用基于词袋的方法具有两大缺陷, 既无法准确知道聚类的数目, 也不能适用于结构复杂(比如呈流形结构)的形状空间. 为此, 本文采用两大方法加以改进, 其一利用有流形聚类功能的决策图方法取代K-means, 其二使用核函数更加有效地衡量三维形状之间的差异. 在SHREC2010库和SHREC2011库上的大量实验结果表明, 两种技巧的有机结合使聚类的精确度和效率得到了显著的提升. |
| Abstract: | In the unsupervised clustering problem of 3D models, the Bag-of-Word based approach has been identified with two major drawbacks: being unable to predict exactly the number of clusters, and unable to apply to a complicated shape space (e.g., with a manifold structure). Therefore, we propose two improvement techniques in this paper. Firstly, the decision graph technique, with a super ability in manifold clustering, is used as a substitute for K-means. Secondly, the difference between shapes is calculated using a combined kernel function, rather than measured in the original signature space. Extensive experimental results on SHREC2010 and SHREC2011 show that the coherent combination of the couple of techniques greatly improves clustering accuracy and efficiency. |
| 参考文献 /References: | [1] Ip C Y, Regli W C. A 3D object classifier for discriminating manufacturing processes[J]. Computers & Graphics, 2006, 30(6):903-916. [2] 王小凤, 周蓬勃, 耿国华, 等. 三维模型自动语义标注技术研究[EB/OL]. [2017-05-12]. http://xueshu.baidu.com/s?wd=paperuri%3A%28bb377e34a54d58de96476d51eb4396af%29&filter=sc_long_sign&tn=SE_xueshusource_2kduw22v&sc_vurl=http%3A%2F%2Fwww.paper.edu.cn%2Fhtml%2Freleasepaper%2F2012%2F02%2F653%2F&ie=utf-8&sc_us=16709243031499231572 2012. [3] 石源, 莫蓉, 常智勇, 等. 基于聚类的模型数据集可视化与检索[J]. 计算机辅助设计与图形学学报, 2010, 22 (11):1918-1924. [4] McWherter D, Peabody M, Shokoufandeh A C, et al. Database techniques for archival of solid models[C]. Symposium on Solid Modeling and Applications, 2001:78-87. [5] 张媛媛. 一种基于非参数贝叶斯模型的聚类算法[J]. 宁波大学学报(理工版), 2013, 26(4):24-28. [6] Li Z M, Wang D Q, Li B, et al. 3D model classification using salient features for content representation[C]. International Conference on Natural Computation, 2010:3541-3545. [7] Heyer L J, Kruglyak S. Exploring expression data: Identification and analysis of coexpressed genes[J]. Genome Research, 1999, 9(11):1106-1115. [8] Alex R, Alessandro L. Machine learning. Clustering by fast search and find of density peaks[J]. Science, 2014, 344(6191):1492-1496. [9] Shapira L, Shamir A, Cohen O D. Consistent mesh partitioning and skeletonisation using the shape diameter function[J]. The Visual Computer, 2008, 24(4):249-259. [10] Gal R, Cohen O D. Salient geometric features for partial shape matching and similarity[J]. ACM Transactions on Graphics, 2006, 25(1):130-150. [11] Bronstein M M, Kokkinos I. Scale-invariant heat kernel signatures for non-rigid shape recognition[C]. Computer Vision and Pattern Recognition, 2010:1704-1711. [12] Aubry M, Schlickewei U, Cremers D. The wave kernel signature: A quantum mechanical approach to shape analysis[C]. IEEE International Conference on Computer Vision Workshops, 2012:626-633. [13] 易雁飞. 基于K-Means聚类的数据分析[J]. 现代制造技术与装备, 2017(4):8-9. [14] 缪元武. 基于层次聚类的数据分析[D]. 合肥: 安徽大学, 2013. [15] 刘紫涵, 吴鹏海, 吴艳兰. 三种谱聚类算法及其应用研究[J]. 计算机应用研究, 2017, 34(4):1026-1031. [16] Jayanti S, Kalyanaraman Y, Ramani K. Shape-based clustering for 3D CAD objects: A comparative study of effectiveness[J]. Computer-Aided Design, 2009, 41(12):999-1007. [17] 舒振宇, 祁成武, 辛士庆, 等. 基于密度峰值的三维模型无监督分类算法[J]. 计算机辅助设计与图形学学报, 2016, 28(12):2142-2150. |
| 备注/Memo: | 收稿日期: 2017－10－17. 宁波大学学报（理工版）网址: http://journallg.nbu.edu.cn/基金项目: 国家自然科学基金（61772016, 61332015）; 浙江省自然科学基金（LY17F020018）; 宁波市自然科学基金（2017A610115）; 浙江大学国家重点实验室开放基金（A1702）.第一作者: 徐欣（1995－）, 女, 安徽合肥人, 在读硕士研究生, 主要研究方向: 人工智能下的三维模型. E-mail: 765641612@qq.com*通信作者: 陈双敏（1982－）, 女, 浙江宁海人, 讲师, 主要研究方向: 计算机图形学、机器学习. E-mail: chenshuangmin@nbu.edu.cn 宁波大学学报（理工版）网址：http://journallg.nbu.edu.cn/ |
