The aim in this paper is to develop a method for clustering together image views of the same object class. Local invariant feature methods, such as SIFT, have been proven effective for image clustering. However, they have made either relatively little use or too complex use of geometric constraints and are confounded when the detected features are superabundant. Here we make two contributions aimed at overcoming these problems. First, we rank the SIFT points (R-SIFT) using visual saliency. Second, we use the reduced set of R-SIFT features to construct a specific hyper graph (CSHG) model of holistic-structure. Based on the CSHG model, a two stage clustering method is proposed. In which, images are clustered according to the pairwise similarity of the graphs, which is a combination of the traditional similarity of local invariant feature vectors and the geometric similarity between two graphs. This method comprehensively utilizes both SIFT and geometric constraints, and hence combines both global and local information. Experiments reveal that the method gives excellent clustering performance.


Visual Saliency Family Tree Probabilistic Latent Semantic Analysis Cluster Rate Noise Free Image 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Xia, S.P., Hancock, E.R.: 3D Object Recognition Using Hyper-Graphs and Ranked Local Invariant Features. In: S+SSPR 2008 (to appear, 2008)Google Scholar
  2. 2.
    Bosch, A., Zisserman, A., Muoz, X.: Scene Classification Using a Hybrid Generative/Discriminative Approach. IEEE Trans. PAMI 30(4), 1–16 (2008)CrossRefGoogle Scholar
  3. 3.
    Lowe, D.G.: Distinctive image features from scale-invariant key points. IJCV 60(2), 91–110 (2004)CrossRefGoogle Scholar
  4. 4.
    Elazary, L., Itti, L.: Interesting objects are visually salient. Journal of Vision 8(3), 1–15 (2008)CrossRefGoogle Scholar
  5. 5.
    Li, F.F., Perona, P.: A Bayesian hierarchical model for learning natural scene categories. CVPR 2, 524–531 (2005)Google Scholar
  6. 6.
    Bay, H., Tuytelaars, T., Gool, L.V.: SURF: Speeded Up Robust Features. In: Leonardis, A., Bischof, H., Pinz, A. (eds.) ECCV 2006, Part I. LNCS, vol. 3951, pp. 404–417. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  7. 7.
    Zhang, J., Marszablek, M., Lazebnik, S., Schmid, C.: Local Features and Kernels for Classification of Texture and Object Categories. IJCV 73(2), 213–238 (2007)CrossRefGoogle Scholar
  8. 8.
    Yan, K., Sukthankar, R.: PCA-SIFT: A more distinctive representation for local image descriptors. CVPR 2, 506–513 (2004)Google Scholar
  9. 9.
    Sivic, J., Zisserman, A.: VideoGoogle: A text retrieval approach to object matching in videos. ICCV 2, 1470–1477 (2003)Google Scholar
  10. 10.
    Nigam, K., Lafferty, J., Mccallum, A.: Using maximum entropy for text classification. In: IJCAI Workshop on Machine Learning for Information Filtering, pp. 61–67 (1999)Google Scholar
  11. 11.
    Sudderth, E.B., Torralba, A., Freeman, W.T., Willsky, A.S.: Describing Visual Scenes Using Transformed Objects and Parts. Int. J. Comput. Vis. 77, 291–330 (2008)CrossRefGoogle Scholar
  12. 12.
    Torsello, A., Hancock, E.R.: Graph Embedding using Tree Edit Union. Pattern Recognition 40, 1393–1405 (2007)CrossRefzbMATHGoogle Scholar
  13. 13.
    Bonev, B., Escolano, F., Lozano, M.A., Suau, P., Cazorla, M.A., Aguilar, W.: Constellations and the Unsupervised Learning of Graphs. In: Escolano, F., Vento, M. (eds.) GbRPR 2007. LNCS, vol. 4538, pp. 340–350. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  14. 14.
    Rizzi, S.: Genetic operators for hierarchical graph clustering. Pattern Recognition Letters 19, 1293–1300 (1998)CrossRefzbMATHGoogle Scholar
  15. 15.
    Segen, J.: Learning graph models of shape. In: Laird, J. (ed.) Proceedings of the Fifth International Conference on Machine Learning, pp. 29–35 (1988)Google Scholar
  16. 16.
    Torsello, A., Hancock, E.R.: Learning shape-classes using a mixture of tree-unions. IEEE Trans. PAMI 28(6), 954–967 (2006)CrossRefGoogle Scholar
  17. 17.
    Torsello, A., Robles-Kelly, A., Hancock, E.R.: Discovering Shape Classes using Tree Edit Distance and Pairwise Clustering. International Journal of Computer Vision 72, 259–285 (2007)CrossRefGoogle Scholar
  18. 18.
    Garey, M., Johnson, D.: Computers and intractability: A guide to the theory of NP-completeness. W.H. Freeman and Company, New York (1979)zbMATHGoogle Scholar
  19. 19.
    Günter, S., Bunke, H.: Self-organizing map for clustering in the graph domain. Pattern Recognition Letters 23, 401–417 (2002)CrossRefzbMATHGoogle Scholar
  20. 20.
    Lozano, M.A., Escolano, F.: Em algorithm for clustering an ensemble of graphs with comb matching. In: Rangarajan, A., Figueiredo, M.A.T., Zerubia, J. (eds.) EMMCVPR 2003. LNCS, vol. 2683, pp. 52–67. Springer, Heidelberg (2003)CrossRefGoogle Scholar
  21. 21.
    Luo, B., Wilson, R.C., Hancock, E.R.: Spectral embedding of graphs. Pattern Recognition 36(10), 2213–2223 (2003)CrossRefzbMATHGoogle Scholar
  22. 22.
    Jain, B.J., Wysotzki, F.: Central Clustering of Attributed Graphs. Machine Learning 56, 169–207 (2004)CrossRefzbMATHGoogle Scholar
  23. 23.
    Hofmann, T., Buhmann, J.M.: Pairwise data clustering by deterministic annealing. IEEE Transactions on Pattern Analysis and Machine Intelligence 19(1), 1–14 (1997)CrossRefGoogle Scholar
  24. 24.
    Xia, S.P., Liu, J.J., Yuan, Z.T., Yu, H., Zhang, L.F., Yu, W.X.: Theory and Algorithm of Machine Learning Based on RSOM Tree Model. ACTA Electronica sinica 33(5), 937–944 (2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Shengping Xia
    • 1
    • 2
  • Edwin R. Hancock
    • 2
  1. 1.ATR Lab, School of Electronic Science and EngineeringNational University of Defense TechnologyChangshaP.R. China
  2. 2.Department of Computer ScienceUniversity of YorkYorkUK

Personalised recommendations