• DocumentCode
    441927
  • Title

    Knowledge-based hierarchical sketch understanding

  • Author

    Zheng, Wen-Tao ; SUN, ZHENG-XING

  • Author_Institution
    State Key Lab for Novel Software Technol., Nanjing Univ., China
  • Volume
    5
  • fYear
    2005
  • fDate
    18-21 Aug. 2005
  • Firstpage
    2838
  • Abstract
    This article presents a knowledge-based framework for sketch understanding. Compared to traditional sketch understanding systems, it has two significant features: (1) knowledge-based. This makes the system independent of applied domain. Users can easily design an ontology-based domain knowledge base to guide the understanding process. (2) hierarchical approach. The sketch understanding process is divided into three phases: token recognition, salient symbol recognition and integral recognition. During all these phases, domain knowledge is used as a guider. In addition, a domain knowledge base of UML class diagram is presented as a case study of this framework, in which the details of how to design ontology-based knowledge base and how to implement a hierarchical recognition are given. Some UML class diagram sketches are used as experiments, which produce satisfactory results.
  • Keywords
    Unified Modeling Language; handwriting recognition; knowledge based systems; ontologies (artificial intelligence); UML class diagram; hierarchical recognition; integral recognition; knowledge-based hierarchical sketch understanding; ontology-based domain knowledge base; salient symbol recognition; token recognition; Automatic speech recognition; Handwriting recognition; Hardware; Learning systems; Ontologies; Shape; Speech recognition; Sun; Unified modeling language; User interfaces; Knowledge base; Sketch understanding; ontology;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
  • Conference_Location
    Guangzhou, China
  • Print_ISBN
    0-7803-9091-1
  • Type

    conf

  • DOI
    10.1109/ICMLC.2005.1527426
  • Filename
    1527426