Classifier Combination Using Random Walks on the Space of Concepts

  • Jorge Sánchez
  • Javier Redolfi
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7441)


We propose a novel approach for the combination of classifiers based on two commonly adopted strategies in multiclass classification: one-vs-all and one-vs-one. The method relies on establishing the relevance of nodes in a graph defined in the space of concepts. Following a similar approach as in the ranking of websites, the relative strength of the nodes is given by the stationary distribution of a Markov chain defined on that graph. The proposed approach do not requires the base classifiers to provide calibrated probabilities. Experiments on the challenging problem of multiclass image classification show the potentiality of our approach.


multiclass classification random walks image classification Fisher vectors 


  1. 1.
  2. 2.
    Çakir, F., Güdükbay, U., Ulusoy, Ö.: Nearest-neighbor based metric functions for indoor scene recognition. Computer Vision and Image Understanding 115(11), 1483–1492 (2011)CrossRefGoogle Scholar
  3. 3.
    Everingham, M., Van Gool, L., Williams, C.K.I., Winn, J., Zisserman, A.: The PASCAL Visual Object Classes Challenge 2007 (VOC 2007) Results,
  4. 4.
    Galar, M., Fernández, A., Tartas, E.B., Sola, H.B., Herrera, F.: An overview of ensemble methods for binary classifiers in multi-class problems: Experimental study on one-vs-one and one-vs-all schemes. Pattern Recognition 44(8), 1761–1776 (2011)CrossRefGoogle Scholar
  5. 5.
    García-Pedrajas, N., Ortiz-Boyer, D.: Improving multiclass pattern recognition by the combination of two strategies. IEEE Tr. on Pattern Analysis and Machine Intelligence 28(6), 1001–1006 (2006)CrossRefGoogle Scholar
  6. 6.
    Haveliwala, T.H.: Topic-sensitive pagerank: A context-sensitive ranking algorithm for web search. IEEE Tr. on Knowledge and Data Eng. 15(4), 784–796 (2003)CrossRefGoogle Scholar
  7. 7.
    Langville, A.N., Meyer, C.D.: Deeper inside PageRank. Internet Mathematics 1(3), 335–400 (2004)MathSciNetzbMATHCrossRefGoogle Scholar
  8. 8.
    Li, L.-J., Su, H., Xing, E.P., Li, F.-F.: Object bank: A high-level image representation for scene classification & semantic feature sparsification. In: Proc. NIPS (2010)Google Scholar
  9. 9.
    Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Intl. Jrnl. on Computer Vision 60(2) (2004)Google Scholar
  10. 10.
    Page, L., Brin, S., Motwani, R., Winograd, T.: The pagerank citation ranking: Bringing order to the web. Technical report, Stanford InfoLab (1999)Google Scholar
  11. 11.
    Perronnin, F., Dance, C.: Fisher kernels on visual vocabularies for image categorization. In: Proc. CVPR (2007)Google Scholar
  12. 12.
    Perronnin, F., Sánchez, J., Mensink, T.: Improving the Fisher Kernel for Large-Scale Image Classification. In: Daniilidis, K., Maragos, P., Paragios, N. (eds.) ECCV 2010, Part IV. LNCS, vol. 6314, pp. 143–156. Springer, Heidelberg (2010)CrossRefGoogle Scholar
  13. 13.
    Quattoni, A., Torralba, A.: Recognizing indoor scenes. In: Proc. CVPR (2009)Google Scholar
  14. 14.
    Reid, S.R.: Model Combination in Multiclass Classification. Ph.D. thesis, Univ. of Colorado (2010)Google Scholar
  15. 15.
    Rifkin, R.M., Klautau, A.: In: defense of one-vs-all classification. Jrnl. of Machine Learning Research 5, 101–141 (2004)MathSciNetzbMATHGoogle Scholar
  16. 16.
    Wu, T.F., Lin, C.J., Weng, R.C.: Probability estimates for multi-class classification by pairwise coupling. Jrnl. of Machine Learning Research 5, 975–1005 (2004)MathSciNetzbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Jorge Sánchez
    • 1
    • 2
  • Javier Redolfi
    • 2
  1. 1.CIEM-CONICETUniversidad Nacional de CórdobaArgentina
  2. 2.CIIIUniversidad Tecnológica Nacional, Fac. Reg. CórdobaArgentina

Personalised recommendations