基于核主成分分析的三维模型分类算法3D Model Classification Algorithm Based on Kernel Principal Component Analysis
王鹏飞,舒振宇,于欣
摘要(Abstract)：
针对三维模型的分类问题,提出了一种基于核主成分分析(Kernel-Principal Components Analysis,K-PCA)的三维模型分类算法。该算法首先选择形状直径函数(Shape Diameter Function,SDF)作为特征描述符来提取三维模型的特征向量;然后使用核函数将原始特征向量映射到高维空间中并在该空间上进行PCA得到新的特征向量;最后使用KNN算法并计算未知模型与已知类别的k个模型之间的l2范数以实现模型的分类,确定未知模型的类别。实验结果表明,该算法能够很好的识别三维模型的几何特征,能准确的区分不同类别的三维模型,具有较高的分类准确率。
关键词(KeyWords)： 核主成分分析;三维模型;K近邻;分类;形状直径函数
基金项目(Foundation): 国家自然科学基金(11226328,61374096);; 浙江省自然科学基金(LY13F020018);; 宁波市自然科学基金(2012A610068)
作者(Author): 王鹏飞,舒振宇,于欣
参考文献(References)：
- [1]ANKERST M,KASTENMüLLER G,KRIEGEL H P,et al.3D shape histograms for similarity search and classification in spatial databases[C]∥Advances in Spatial Databases.Springer Berlin Heidelberg,1999:207-226.
- [2]ELAD M,TAL A,AR S.Content based retrieval of VRML objects—an iterative and interactive approach[M].Springer Vienna,2002,18(3):107-118.
- [3]ZHANG C,CHEN T.Efficient feature extraction for 2D/3D objects in mesh representation[C]∥Image Processing,2001.Proceedings of 2001 International Conference on IEEE,2001,3:935-938.
- [4]HUBER D,KAPURIA A,DONAMUKKALA R,et al.Parts-based 3d object classification[C]∥Proceedings of the 2004 IEEE Computer Society Conference on IEEE,2004,2:Ⅱ-82-Ⅱ-89.
- [5]CHUA C S,JARVIS R.Point signatures:A new representation for 3d object recognition[J].International Journal of Computer Vision,1997,25(1):63-85.
- [6]KAZHDAN M,FUNKHOUSER T,RUSINKIEWICZ S.Rotation invariant spherical harmonic representation of 3 D shape descriptors[C]∥Symposium on geometry processing.2003.
- [7]OSADA R,FUNKHOUSER T,CHAZELLE B,et al.Matching 3D models with shape distributions[C]∥Shape Modeling and Applications,SMI 2001 International Conference on IEEE,2001:154-166.
- [8]刘斌,上官宁,江开勇.三角网格模型分类分析[J].华侨大学学报:自然科学版,2009,30(6):606-609.
- [9]WANG Y,LU T,GAO R,et al.3D model comparison through kernel density matching[C]∥Pattern Recognition(ICPR),201020th International Conference on IEEE,2010:3159-3162.
- [10]冯立颖赵静杨莹.基于轮廓特征点的三维模型相似性分类算法[J].计算机应用,2010,30(4):914-916.
- [11]陈俊英,王羡慧,方亚萍.基于RBF神经网络集成的三维模型分类和检索[J].图学学报,2013,34(2):26-30.
- [12]高恩阳,刘伟军,王天然,等.基于谱特征和深度信任网络的三维模型分类[J].机械设计与制造,2013(3):250-252.
- [13]付小君,郭鹏江,郭竞,等.统计特征和Markov模型在三维模型分类中的应用[J].Computer Engineering and Applications,2011,47(4):157-159.
- [14]ZHU K P,WONG Y S,LU W F,et al.3D CAD model matching from 2D local invariant features[J].Computers in industry,2010,61(5):432-439.
- [15]CHANG J S,SHIH A C C,LIN H Y,et al.3-D mesh representation and retrieval using Isomap manifold[C]∥Multimedia Signal Processing,2008 IEEE 10th Workshop on IEEE,2008:541-546.
- [16]SHAPIRA L,SHAMIR A,COHEN-OR D.Consistent mesh partitioning and skeletonisation using the shape diameter function[J].The Visual Computer,2008,24(4):249-259.
- [17]张秀琴,陈立潮,潘理虎,等.基于DCT和分块2D2PCA的人脸识别[J].太原科技大学学报,2014,35(5):333-338.
- [18]Shape Analysis Research Project.SHREC'11 3D Models Dataset[EB/OL].[2014-10-09].http:∥www.itl.nist.gov/iad/vug/sharp/contest/2011/Non.
- [19]LIU Y G,YOU Z S,CAO L P.A novel and quick SVM-based multi-class classifier[J].Pattern Recognition,2006,39(11):2258-2264.
- [20]BRONSTEIN M M,KOKKINOS I.Scale-invariant heat kernel signatures for non-rigid shape recognition[C]∥Proceedings of2010 IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Los Alamitos:IEEE Computer Society Press,2010:1704-1711.