Objective evaluation of the discriminant power of features in an HMM-based word recognition system

  • A. El-Yacoubi
  • M. Gilloux
  • R. Sabourin
  • C. Y. Suen
Invited Talks
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1339)


This paper describes an elegant method for evaluating the discriminant power of features in the framework of an HMM-based word recognition system. This method employs statistical indicators, entropy and perplexity, to quantify the capability of each feature to discriminate between classes without resorting to the result of the recognition phase. The HMMs and the Viterbi algorithm are used as powerful tools to automatically deduce the probabilities required to compute the above mentioned quantities.


Handwritten Word Recognition Feature Extraction Discriminant Power of Features Hidden Markov Models 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [Bahl83]
    Bahl L., Jelinek F., Mercer R. “A maximum likelihood approach to speech recognition”, IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI) 5:179–190, 1983.Google Scholar
  2. [Caesar94]
    Caesar T, Gloger J.M., Kaltenmeier A., Mandler E., “Handwritten Word Recognition Using Statistics,” The Institution of Electrical Engineers. Printed and published by the IEE, Savoy Place, London WC2R OBL, UK, 1994.Google Scholar
  3. [Chen94]
    Chen M.Y, Kundu A., Zhou J., “Off Line Handwritten Word Recognition Using a Hidden Markov Model Type Stochastic Network,” IEEE, TPAMI, Vol. 16, No. 5, May 1994, pp. 481–496.Google Scholar
  4. [Cohen94]
    Cohen E., Hull J.J., Srihari S.N., “Control Structure for Interpreting Handwritten Addresses,” PAMI, Vol. 16, NO. 10, October 1994.Google Scholar
  5. [Cox88]
    Cox S.J., “Hidden Markov Models for Automatic Speech Recognition: Theory and Application,” Br Telecom Technol J Vol. 6 No. 2, April 1988.Google Scholar
  6. [ElYacoubi94]
    El-Yacoubi A., Bertille J.M., Gilloux M., “Towards a more effective handwritten word recognition system,” Proceedings of the Fourth International Workshop on Frontiers in Handwriting Recognition (IWFHR-IV), pp. 378–385, Taiwan, December 1994.Google Scholar
  7. [EIYacoubi95]
    El-Yacoubi A., Bertille J.M., Gilloux M., “Conjoined Location and Recognition of Street Names Within a Postal Address Delivery Line,” ICDAR'95, pp. 1024–1027, Vol. 2, Montréal, August 1995.Google Scholar
  8. [ElYacoubi96]
    A. El-Yacoubi. “Modélisation Markovienne de l'écriture manuscrite. Application à la reconnaissance des adresses postales,” PhD thesis, Université de Rennes I, France, 1996.Google Scholar
  9. [Forney73]
    Forney G.D., “The Viterbi Algorithm,” Proceedings of the IEEE, Vol. 61, NO. 3, March 1973.Google Scholar
  10. [Gillies92]
    Gillies A.M., “Cursive Word Recognition Using Hidden Markov Models,” Proc. of the 5th US-Postal Service Advanced Technology Conf., pp. 557–562, Nov 30–Dec 2, 1992.Google Scholar
  11. [Gilloux92]
    Gilloux M., Leroux M. “Recognition of cursive script amounts on postal cheques,” Proc. of the 5th USPS Advanced Technology Conf., pp. 545–556, 1992.Google Scholar
  12. [Gilloux95a]
    Gilloux M., Leroux M., Bertille J-M., “Strategies for Cursive Script Recognition Using Hidden Markov Models,” Machine Vision and Applications, Vol. 8, pp. 197–205,1995.Google Scholar
  13. [Gorski94]
    Gorski N.D., “Experiments with Handwriting Recognition using Holographic Representation of Line Images,” Pattern Recognition Letters 15, pp. 853–859, 1995.Google Scholar
  14. [Kim95]
    Kim G., Govindaraju V., “Handwritten Word Recognition for Real-Time Applications,” International Conference on Document Analysis and Recognition (ICDAR), Vol. 1, pp. 24–27, August 14–16,1995.Google Scholar
  15. [Kim97]
    Kim G., Govindaraju V, “A Lexicon Driven Approach to Handwritten Word Recognition for Real-Time Applications,” PAMI, Vol. 19, No. 4, April 1997, pp. 366–379.Google Scholar
  16. [Knerr96]
    Knerr S., Baret O., Price D., Simon J.C., “The A2iA Recognition System for Handwritten Checks,” Proceedings of Document Analysis Systems, Malvern, Pennsylvania, October 14–16, 1996, pp. 431–494.Google Scholar
  17. [Lee91]
    Lee K.F, Hon H.W, Hwang M.Y, Huang X., “Speech Recognition Using Hidden Markov Models: A CMU Perspective,” Elsevier Science Publishers B.V. (North-Holland), 1991, pp. 497–508.Google Scholar
  18. [Leroux91]
    Leroux M., “Reconnaissance de Textes Manuscrits à Vocabulaire Limité avec Application à la Lecture Automatique des Chèques,” PhD thesis, Université de Rouen, 1991.Google Scholar
  19. [Paquet93]
    Paquet T., Lecourtier Y., “Recognition of Handwritten Sentences Using A Restricted Lexicon,” Pattern Recognition, Vol. 26, NO. 3, pp. 391–407, 1993.Google Scholar
  20. [Poritz88]
    Poritz A.B., “Hidden Markov Models: A Guided Tour,” Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'88), pp. 7–13, 1988.Google Scholar
  21. [Rabiner93]
    Rabiner L., Juang B-H, “Fundamentals of Speech Recognition,” Prentice Hall Signal Processing Series, pp. 382–384, 1993.Google Scholar
  22. [Rabiner89]
    Rabiner L.R., “A tutorial on Hidden Markov Models and Selected Applications in Speech Recognition,” Proc. IEEE, 77 (2): 257–286, February 1989.Google Scholar
  23. [Shrihari95]
    Shrihari S.N, Shin Y-C, Ramanaprasad V, Lee D-S, “Name and Address Block Reader System for Tax Form Processing,” International Conference on Document Analysis and Recognition (ICDAR), Vol. 1, pp. 5–10, August 14–16, 1995.Google Scholar
  24. [Simon92]
    Simon J.C., Baret O., “Cursive Word Recognition,” From Pixels to Features III. Frontiers in Handwriting Recognition, S. Impedovo and J.C. Simon Editors (North-Holland), pp. 241–260, 1992.Google Scholar
  25. [Suen96]
    Suen C.Y, Lam L., Guillevic D., Strathy N.W., Chenet M., Said J.N., Fan R., “Bank Check Processing System,” International Journal of Imaging Systems and Technology, Vol. 7, 392–403 (1996).Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • A. El-Yacoubi
    • 1
    • 2
  • M. Gilloux
    • 2
  • R. Sabourin
    • 1
    • 3
  • C. Y. Suen
    • 1
  1. 1.Centre for Pattern Recognition and Machine Intelligence Department of Computer ScienceConcordia UniversityMontréalCanada
  2. 2.Service de Recherche Technique de La Poste Department ReconnaissanceModélisation Optimisation (RMO)Nantes Cedex 02France
  3. 3.Ecole de Technologie SupérieureLaboratoire d'imagerie, de vision et d'intelligence artificielle (LIVIA)Notre-Dame OuestCanada

Personalised recommendations