Foundations of a Multi-way Spectral Clustering Framework for Hybrid Linear Modeling

被引:48
作者
Chen, Guangliang [1 ]
Lerman, Gilad [1 ]
机构
[1] Univ Minnesota, Dept Math, Minneapolis, MN 55455 USA
基金
美国国家科学基金会;
关键词
Hybrid linear modeling; d-flats clustering; Multi-way clustering; Spectral clustering; Polar curvature; Perturbation analysis; Concentration inequalities; SEGMENTATION;
D O I
10.1007/s10208-009-9043-7
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
The problem of Hybrid Linear Modeling (HLM) is to model and segment data using a mixture of affine subspaces. Different strategies have been proposed to solve this problem, however, rigorous analysis justifying their performance is missing. This paper suggests the Theoretical Spectral Curvature Clustering (TSCC) algorithm for solving the HLM problem and provides careful analysis to justify it. The TSCC algorithm is practically a combination of Govindu's multi-way spectral clustering framework (CVPR 2005) and Ng et al.'s spectral clustering algorithm (NIPS 2001). The main result of this paper states that if the given data is sampled from a mixture of distributions concentrated around affine subspaces, then with high sampling probability the TSCC algorithm segments well the different underlying clusters. The goodness of clustering depends on the within-cluster errors, the between-clusters interaction, and a tuning parameter applied by TSCC. The proof also provides new insights for the analysis of Ng et al. (NIPS 2001).
引用
收藏
页码:517 / 558
页数:42
相关论文
共 43 条
  • [1] Agarwal S, 2005, PROC CVPR IEEE, P838
  • [2] Agarwal S., 2006, ICML, P17, DOI DOI 10.1145/1143844.1143847
  • [3] [Anonymous], COMP VIS PATT REC WO
  • [4] ARIASCASTRO E, 2005, IEEE T INF THEORY, V51
  • [5] Algorithm 862: MATLAB tensor classes for fast algorithm prototyping
    Bader, Brett W.
    Kolda, Tamara G.
    [J]. ACM TRANSACTIONS ON MATHEMATICAL SOFTWARE, 2006, 32 (04): : 635 - 653
  • [6] k-plane clustering
    Bradley, PS
    Mangasarian, OL
    [J]. JOURNAL OF GLOBAL OPTIMIZATION, 2000, 16 (01) : 23 - 32
  • [7] Brand M., 2003, PMLR
  • [8] Spectral Curvature Clustering (SCC)
    Chen, Guangliang
    Lerman, Gilad
    [J]. INTERNATIONAL JOURNAL OF COMPUTER VISION, 2009, 81 (03) : 317 - 330
  • [9] A multibody factorization method for independently moving objects
    Costeira, JP
    Kanade, T
    [J]. INTERNATIONAL JOURNAL OF COMPUTER VISION, 1998, 29 (03) : 159 - 179
  • [10] A multilinear singular value decomposition
    De Lathauwer, L
    De Moor, B
    Vandewalle, J
    [J]. SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2000, 21 (04) : 1253 - 1278