In Pattern Recognition, there are problems where distinct representations can be obtained for the same pattern, and depending on the type of classifiers (statistical or structural) one type of representation is preferred versus the others. In the last years, different approaches to combining classifiers have been proposed to improve the performance of individual classifiers. However, few works investigated the use of structured pattern representations. In this paper combination of classifiers has been applied using tree pattern representation in combination with strings and vectors for a handwritten character classification task. In order to save computational cost, some proposals based on the use of both embedding structured data and refine and filter framework are provided.


combining classifiers tree representation string representation edit distance k-NN rule 


  1. 1.
    Kuncheva, L.: Combining Pattern Classifiers: methods and algorithms. Wiley, Chichester (2004)CrossRefzbMATHGoogle Scholar
  2. 2.
    Duin, R.: The combining classifier: to train or not to train? In: Proceedings of the International Conference on Pattern Recognition ICPR 2002, Quebec Canada, vol. II, pp. 765–770 (2002)Google Scholar
  3. 3.
    Neuhaus, M., Bunke, H.: Edit distance-based kernel functions for structural pattern classification. Pattern Recognition 39(10), 1852–1863 (2006)CrossRefzbMATHGoogle Scholar
  4. 4.
    Spillmann, B., Neuhaus, M., Bunke, H., Pekalska, E., Duin, R.: Transforming strings to vector spaces using prototype selection. In: Yeung, D.-Y., Kwok, J.T., Fred, A., Roli, F., de Ridder, D. (eds.) SSPR 2006 and SPR 2006. LNCS, vol. 4109, pp. 287–296. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  5. 5.
    Grother, P.: Handprinted forms and character database. In: NIST Special Database 19 (2001)Google Scholar
  6. 6.
    Carrasco, R., Forcada, M.: A note on the Nagendraprasad-Wnag-Gupta thinning algorithm. Patter Recognition Letters 16, 539–541 (2003)CrossRefzbMATHGoogle Scholar
  7. 7.
    Rico, J., Micó, L.: Comparison of aesa and laesa search algorithms using string and tree edit distances. Pattern Recognition Letters 24(9), 1427–1436 (2003)zbMATHGoogle Scholar
  8. 8.
    Zhang, K., Shasha, D.: Simple fast algorithms for the editing distance between trees and related problems. SIAM J. Comput. 18, 1245–1262 (1989)MathSciNetCrossRefzbMATHGoogle Scholar
  9. 9.
    Wagner, R., Fischer, M.: The string-to-string correction problem. Journal of the Association for Computing Machinery 211, 168–173 (1974)MathSciNetCrossRefzbMATHGoogle Scholar
  10. 10.
    Kittler, J., Hatef, M., Duin, R., Matas, J.: On combining classifiers. IEEE Trans. on Pattern Analysis and Machine Intelligence 20, 226–239 (1998)CrossRefGoogle Scholar
  11. 11.
    Hjaltason, G., Samet, H.: Properties of embedding methods for similarity searching in metric spaces. IEEE Trans. on Pattern Analysis and Machine Intelligence 25, 530–549 (2003)CrossRefGoogle Scholar
  12. 12.
    Duin, R.P.W., Tax, D.M.J.: Classifier conditional posterior probabilities. In: Amin, A., Pudil, P., Dori, D. (eds.) SPR 1998 and SSPR 1998. LNCS, vol. 1451, pp. 611–619. Springer, Heidelberg (1998)CrossRefGoogle Scholar
  13. 13.
    Arlandis, J., Perez-Cortes, J.C., Cano, J.: Rejection strategies and confidence measures for a k-nn classifier in an ocr task. In: 16th International Conference on Pattern Recognition ICPR 2002, pp. 576–579. IEEE, Los Alamitos (2002)Google Scholar
  14. 14.
    Athitsos, V., Alon, J., Sclaroff, S.: Efficient nearest neighbor classification using a cascade of approximate similarity measures. In: Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2005), vol. 1, pp. 486–493 (2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • Raisa Socorro
    • 1
  • Luisa Micó
    • 2
  1. 1.Instituto Superior Politécnico Jose Antonio EchevarríaLa HabanaCuba
  2. 2.Departamento de Lenguajes y Sistemas InformáticosUniversidad de AlicanteAlicanteSpain

Personalised recommendations