Knowledge-based approach for adaptive recognition of drawings

  • Shin-ichirou Okazaki
  • Yoshitake Tsuji
Knowledge-Based Methods
Part of the Lecture Notes in Computer Science book series (LNCS, volume 301)


This paper describes the Knowledge-based Adaptive Recognition System for Drawings (KARD), using a computer and an optical scanner, and its application to chemical structural formula. KARD is designed as a test-bed system for a drawings recognition method with constructional knowledge about drawings concerned. In order to adaptively resolve target variety and distortion variety, KARD adopts an objective pattern unit description of rules and data-flow-like activation control mechanism for the rules. KARD application to chemical structural formula recognition, with knowledge about its syntactic restrictions, proves that KARD can not only recognize them but also detect contradictions due to recognition failures in the processes.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    H.Bunke, “Experience with several methods for the analysis of schematic diagrams”, Proc. of 6th ICPR, pp710–712, 1982.Google Scholar
  2. [2]
    H.Harada, Y.Ishii, “Recognition of freehand drawings in chemical plant engineering”, IEEE workshop on CAPAIDM, pp146–153, 1985.Google Scholar
  3. [3]
    T.Matsuyama, “Knowledge organization and control structure in image understanding”, Proc. of 7th ICPR, pp1118–1127, 1984.Google Scholar
  4. [4]
    T.Kanade, “Model representations and control structures in image understanding”, Proc. of 5th IJCAI, pp1074–1082, 1977.Google Scholar
  5. [5]
    R.A.Brooks, “The ACRONYM model based vision system”, Proc. of 6th IJCAI, pp105–113. 1979.Google Scholar
  6. [6]
    A.M. Nazif, M.D. Levine, “Low level image segmentation. An expert system”, PAMI vol.6 No.5, pp555–577, 1984.Google Scholar
  7. [7]
    D.M. McKeown, Jr., W.A. Harvey, J. McDermott, “Rule-based interpretation of aerial imagery”, PAMI Vol.7 No.5, pp570–585, 1985.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1988

Authors and Affiliations

  • Shin-ichirou Okazaki
    • 1
  • Yoshitake Tsuji
    • 1
  1. 1.C&C Information Technology Research Laboratories, NEC CorporationKawasaki, KanagawaJapan

Personalised recommendations