Measuring the Robustness of Character Shape Coding

  • A. Lawrence Spitz
  • Paul Marks
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1655)

Abstract

Earlier claims of great robustness for the character shape coding process have been largely unsupported.We provide quantitative measures of the sensitivity of the character shape coding process to the text input, production values and image quality and to the complexity of the destination character shape codes. Using this evaluation tool we can tune the character shape coding process in a systematic way and also develop new versions of the shape codes appropriately adapted to particular applications.

Keywords

Text Image Document Image Optical Character Recognition High Frequency Word Character Shape 
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.

Reference

  1. 1.
    A. L. Spitz, “ Generalized line, word and character finding”, Progress in Image Analysis and Processing III. S. Impedovo (ed.), pp. 377–383, World Scientific, (1993).Google Scholar
  2. 2.
    A. L. Spitz, “ Text characterization by connected component transformation”, Proc. SPIE, San Jose, CA, pp. 97–105, (1994).Google Scholar
  3. 3.
    A. L. Spitz, “ Analysis of compressed document images for dominant skew, multiple skew and logotype detection”, Comp. Vision Image Understanding, 70, 3, pp. 321–334, (1998).CrossRefGoogle Scholar
  4. 4.
    J. Esakov, D. P. Lopresti & J. S. Sandberg., “ Classification and distribution of optical character recognition errors”, Proc. SPIE, San Jose, CA, pp. 204–216, (1994).Google Scholar
  5. 5.
    H. S. Baird, “ Document image defect models”, H.S. Baird, H. Bunke, and K. Yamamoto, eds., Structured Document Analysis. New York, Springer-Verlag, pp.546–556, (1992).Google Scholar
  6. 6.
    R. A. Wagner, & M. J. Fischer, “ The String-to-String Correction Problem”, J. Assoc. Computing Machinery, 21, 1, pp. 168–173, (1974).MATHMathSciNetGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • A. Lawrence Spitz
    • 1
  • Paul Marks
    • 1
  1. 1.Document Recognition Technologies, Inc.Palo AltoUSA

Personalised recommendations