古文書解読用電子くずし字辞典を引くための手書き文字認識  [in Japanese] Development of cursive character pattern recognition for accessing a digital dictionary to support decoding of historical documents  [in Japanese]

Search this Article

Author(s)

Abstract

本稿では,電子くずし字辞典に用いるロバストな文字認識の開発について述べる.古文書の翻刻作業を特定の専門家以外でも可能にし,その効率を高めるために,古文書で標準的に用いられるくずし字の辞典を電子化し,翻刻の利便性を向上させることが有効である.我々は,67,739種のくずし字に対する認識システムを開発した.くずし字までいかない通常の字体に対しても現有の認識システムを利用できるようにした.採用手法は,現在の文字パターンに対して一定の認識率を保証しているので,実用に耐えることを期待しているが,現実課題のサンプルパターンが少ないために,定量的な評価は今後の課題とする.

This paper presents robust recognition of cursive character patterns for accessing a digital dictionary to decode historical documents. In order to liberate the task of decoding historical documents from a very limited number of experts and make the task efficient, preparation of a digital dictionary of cursive styles of character patterns commonly used in historical documents is effective. We have implemented a cursive handwriting recognition system for 67,739 categories. Moreover, we made the existing recognition system available for recognizing less cursive handwritten character patterns. Since the employed technologies have been already established for current characters and handwritings, we expect the method to work for the above-mentioned our task. Sample patterns are not enough, however, so that performance evaluation is left for later studies.

Journal

  • IEICE technical report

    IEICE technical report 106(606), 91-96, 2007-03-09

    The Institute of Electronics, Information and Communication Engineers

References:  10

Cited by:  2

Codes

  • NII Article ID (NAID)
    110006248971
  • NII NACSIS-CAT ID (NCID)
    AN10541106
  • Text Lang
    JPN
  • Article Type
    Journal Article
  • ISSN
    09135685
  • NDL Article ID
    8707674
  • NDL Source Classification
    ZN33(科学技術--電気工学・電気機械工業--電子工学・電気通信)
  • NDL Call No.
    Z16-940
  • Data Source
    CJP  CJPref  NDL  NII-ELS 
Page Top