• DocumentCode
    2176741
  • Title

    Faxed document image restoration using gray level representation

  • Author

    Oguro, Masami ; Akiyama, Temo ; Ogura, Kenji

  • Author_Institution
    NTT Human Interface Labs., Kanagawa, Japan
  • Volume
    2
  • fYear
    1997
  • fDate
    18-20 Aug 1997
  • Firstpage
    679
  • Abstract
    The paper describes an image restoration algorithm for degraded faxed images, for example, salt and pepper noise and the partial omission of figures. Removing these types of degradation is paramount to achieving high performance fax OCR systems. Some observations that enable effective restoration of fax degradation are: (I) the contour shapes of figures may be preserved microscopically; (II) degradation characteristics can be estimated by utilizing the distribution of neighboring pixel values; and (III) the ununiformity of sensor sensitivity causes similar degradation in vertical bands. The proposed method based on these observations comprises the following steps: producing gray level images, restoration based on local pixel distribution, and restoration based on vertical specification. The experimental results based on actual faxed documents show the effectiveness of the method
  • Keywords
    document image processing; facsimile; image restoration; optical character recognition; contour shapes; degradation characteristics; degraded faxed images; faxed document image restoration; faxed documents; gray level images; gray level representation; high performance fax OCR systems; image restoration algorithm; local pixel distribution; neighboring pixel values; partial omission; salt and pepper noise; sensor sensitivity; vertical bands; vertical specification; Background noise; Character recognition; Degradation; Facsimile; Image restoration; Image sensors; Noise figure; Optical character recognition software; Pixel; Sensor phenomena and characterization;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Document Analysis and Recognition, 1997., Proceedings of the Fourth International Conference on
  • Conference_Location
    Ulm
  • Print_ISBN
    0-8186-7898-4
  • Type

    conf

  • DOI
    10.1109/ICDAR.1997.620593
  • Filename
    620593