Top-Down Likelihood Word Image Generation Model for Holistic Word Recognition

  • Eiki Ishidera
  • Simon M. Lucas
  • Andrew C. Downton
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 2423)

Abstract

This paper describes a new top-down word image generation model for word recognition. This model can generate a word image with a likelihood based on linguistic knowledge, segmentation and character image. In the recognition process, first, the model generates the word image which approximates an input image best for each of a dictionary of possible words. Next, the model calculates the distance value between the input image and each generated word image. Thus, the proposed method is a type of holistic word recognition method. The effectiveness of the proposed method was evaluated in an experiment using type-written museum archive card images. The difference between a non-holistic method and the proposed method is shown by the evaluation. The small errors accumulate in non-holistic methods during the process carried out, because the non-holistic methods can’t cover the whole word image but only part images extracted by segmentation, and the non-holistic method can’t eliminate the blackpixels intruding in the recognition window from neighboring characters. In the proposed method, we can expect that no such errors will accumulate. Results show that a recognition rate of 99.8% was obtained, compared with only 89.4% for a recently published comparator algorithm.

References

  1. 1.
    A. C. Downton et al.: “Constructing Web-Based Legacy Index Card Archives — Architectual Design Issues and Initial Data Acquisition”, ICDAR’1, 2001.Google Scholar
  2. 2.
    M. Sawaki, N. Hagita: “Text-line Extraction and Character Recognition of Document Headlines with Graphical Designs using Complementary Similarity Measure”, IEEE trans. PAMI, vol. 20, No. 10, pp 1103–1109, 1998.Google Scholar
  3. 3.
    S. M. Lucas et al.: “RobustWord Recognition for Museum Archive Card Indexing”, ICDAR’01, 2001.Google Scholar
  4. 4.
    M. A. Ozdil, F.T.Y. Vural: “Optical Character Recognition without Segmentation”, ICDAR’97, 1997Google Scholar
  5. 5.
    S. Madhvanath, E. Kleinberg, V. Govindaraju: “Holistic Verification of Handwritten Phrases”, IEEE trans. PAMI, vol. 21, No. 12, pp 1344–1356, 1999.Google Scholar
  6. 6.
    S. Madhvanath, V. Govindaraju: “The Role of Holistic Paradigms in Handwritten Word Recognition”, IEEE trans. PAMI, vol. 23, No. 2, pp. 149–164, 2001.Google Scholar
  7. 7.
    Y. Lu, C.L. Tan, W. Huang, L. Fan:“An Approach to Word Image Matching Based on Weighted Hausdor. Distance”, ICDAR’01, 2001.Google Scholar
  8. 8.
    R. Plamondon, S.N. Srihari: “On-Line and O.-Line Handwriting Recognition: A Comprehensive Survey”, IEEE trans. PAMI, vol. 22, No. 1, pp 63–84, 2000.Google Scholar
  9. 9.
    E. Lethelier, M. Leroux, M. Gilloux: “An Automatic Reading System for Handwritten Numeral Amounts on French Checks”, ICDAR’95, 1995Google Scholar
  10. 10.
    M. Gilloux, B. Lemaré, M. Leroux: “A Hybrid Radial Basis Function Network/Hidden Markov Model Handwritten Recognition System”, ICDAR’95, 1995Google Scholar
  11. 11.
    I. Bazzi, C. LaPre, J. Makhoul, C. Raphael, R. Schwartz: “ Omnifont and Unlimited-Vocabulary OCR for English and Arabic”, ICDAR’97,1997.Google Scholar
  12. 12.
    P. Sarkar, G. Nagy: “Style-consistency in isogenous patterns”, ICDAR’01, 2001.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2002

Authors and Affiliations

  • Eiki Ishidera
    • 1
  • Simon M. Lucas
    • 2
  • Andrew C. Downton
    • 3
  1. 1.NEC CorporationMultimedia Research LabsKawasakiJapan
  2. 2.Dept. of Computer ScienceUniversity of EssexColchesterUK
  3. 3.Dept. of Electronic Systems EngineeringUniversity of EssexColchesterUK

Personalised recommendations