On the Behavior of Kernel Mutual Subspace Method

  • Hitoshi Sakano
  • Osamu Yamaguchi
  • Tomokazu Kawahara
  • Seiji Hotta
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6469)


Optimizing the parameters of kernel methods is an unsolved problem. We report an experimental evaluation and a consideration of the parameter dependences of kernel mutual subspace method (KMS). The following KMS parameters are considered: Gaussian kernel parameters, the dimensionalities of dictionary and input subspaces, and the number of canonical angles. We evaluate the recognition accuracies of KMS through experiments performed using the ETH-80 animal database. By searching exhaustively for optimal parameters, we obtain 100% recognition accuracy, and some experimental results suggest relationships between the dimensionality of subspaces and the degrees of freedom for the motion of objects. Such results imply that KMS achieves a high recognition rate for object recognition with optimized parameters.


Training Sample Object Recognition Recognition Rate Recognition Accuracy Kernel Parameter 
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.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Maeda, K.-i., Watanabe, S.: Pattern matching method with local structure. Trans. on IEICE (D) 68-D(3), 345–352 (1985) (Japanese Edition); Cipolla, R., et al. (ed.) recent English version is, Computer Vision: Detection, Recognition and Reconstruction (Studies in Computational Intelligence) Part 5. From the Subspace Methods to the Mutual Subspace Method.SCI. Springer Heidelberg (2010)Google Scholar
  2. 2.
    Yamaguchi, O., Fukui, K., Maeda, K.: Face recognition using temporal image sequence. In: Proc. of IEEE 4th Int’l. Conf. on Face and Gesture Recognition, pp. 318–323 (1998)Google Scholar
  3. 3.
    Sakano, H., Mukawa, N.: Kernel mutual subspace method for robust facial image recognition. In: Proc. of 4th Int’l. Conf. on Knowledge based Engineering System, Brighton, vol. 1, pp. 245–248 (2000)Google Scholar
  4. 4.
    Sakano, H., Mukawa, N., Nakamura, T.: Kernel mutual subspace method and its application for object recognition. Electronics and Communications in Japan E88(6), 45–53 (2005)Google Scholar
  5. 5.
    Sakano, H., Suenaga, T.: Classifiers under continuous observations. In: Caelli, T.M., Amin, A., Duin, R.P.W., Kamel, M.S., de Ridder, D. (eds.) SPR 2002 and SSPR 2002. LNCS, vol. 2396, pp. 798–805. Springer, Heidelberg (2002)CrossRefGoogle Scholar
  6. 6.
    Ichino, M., Sakano, H., Komatsu, N.: Speaker recognition using kernel mutual subspace method. In: Proc. of ICARCV (2004)Google Scholar
  7. 7.
    Zhang, B., Park, J., Ko, H.: Combination of self-organization map and kernel mutual subspace method for video surveillance. Advanced Video and Signal Based Surveillance, 123–128 (2007)Google Scholar
  8. 8.
    Ichino, M., Sakano, H., Komatsu, N.: Text-indicated speaker recognition using kernel mutual subspace method. In: Proc. of ICARCV, Singapore, p. 11027 (2008)Google Scholar
  9. 9.
    Fujimaki, R., Yairi, T., Machida, K.: An approach to spacecraft anomaly detection problem using kernel feature space. In: 11th ACM SIGKDD Int’l Conf. on Knowledge Discovery and Data Mining (KDD 2005), pp. 401–410 (2005)Google Scholar
  10. 10.
    Fukui, K., Yamaguchi, O.: A Theoretical Extension of the Subspace Method and its Application for 3D Object Recognition. IPSJ Transactions on Computer Vision and Image Media 46(SIG 15(CVIM 12)), 21–34 (2005) (in Japanese)Google Scholar
  11. 11.
    Leibe, B., Schiele, B.: Analyzing appearance and contour based methods for object categorization. In: CVPR, pp. 409–415 (2003)Google Scholar
  12. 12.
    Chatelin, F.: Veleurs propres de matrices. Masson, Paris (1988) (in French)Google Scholar
  13. 13.
    Kim, T.-K., Arandjelovic, O., Cipolla, R.: Boosted manifold principal angles for image set-based recognition. Pattern Recognition 40(9), 2475–2484 (2007)CrossRefzbMATHGoogle Scholar
  14. 14.
    Maeda, K., Yamaguchi, O., Fukui, K.: Towards 3-dimensional pattern recognition. In: SSPR/SPR 2004, pp.1061–1068 (2004)Google Scholar
  15. 15.
    Schölkopf, B., Smola, A., Müler, K.R.: Nonlinear component analysis as a kernel eigenvalue problem. Neural Computation 10, 1299–1319 (1998)CrossRefGoogle Scholar
  16. 16.
    Fukui, K., Stenger, B., Yamaguchi, O.: A framework for 3D object recognition using the kernel constrained mutual subspace method. In: Proceedings of Asian Conference on Computer Vision, pp. 315–324 (2006)Google Scholar
  17. 17.
    Fukui, K., Yamaguchi, O.: The kernel orthogonal mutual subspace method and its application to 3D object recognition. In: Yagi, Y., Kang, S.B., Kweon, I.S., Zha, H. (eds.) ACCV 2007, Part II. LNCS, vol. 4844, pp. 467–476. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  18. 18.
    Zhang, J., Marszalek, M., Lazebnik, S., Schmid, C.: Local features and kernels for classification of texture and object categories: A comprehensive study. Int’l J. of Computer Vision 73(2), 213–238 (2007)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Hitoshi Sakano
    • 1
  • Osamu Yamaguchi
    • 2
  • Tomokazu Kawahara
    • 3
  • Seiji Hotta
    • 4
  1. 1.NTT Communication Science Lab.Cityh KyotoJapan
  2. 2.Power and Industrial System R&D CenterToshiba Corporation Power Systems CompanyTokyoJapan
  3. 3.Corporate R&D CenterToshiba CorporationKawasakiJapan
  4. 4.The Graduate School of EngineeringTokyo University of Agriculture and TechnologyTokyoJapan

Personalised recommendations