• DocumentCode
    2481289
  • Title

    Gray-based news video text extraction approach

  • Author

    Xiaoling, Fu ; Hua, Gao

  • Author_Institution
    Inf. Eng. Coll., North China Univ. Of Technol., Beijing, China
  • fYear
    2010
  • fDate
    Nov. 30 2010-Dec. 2 2010
  • Firstpage
    208
  • Lastpage
    211
  • Abstract
    In the news video, As a high-level semantic features, video text play a crucial role for the semantics understanding, video analysis, quick video clips retrieval of news video. We propose a new video character extraction method that is gray-based. It makes full use of gray information of the video image and the news video text features to detect the video frame which contain text information, and uses mathematical statistical method to find the location where text appears, then do the binary processing of the candidate region image, at last uses OCR software package to recognize the text. This article also finish a simple news video retrieval system, and text extraction algorithm is applied to the system. Experiment results on Chinese CCTV news video show the encouraging performance of the proposed algorithm.
  • Keywords
    optical character recognition; software packages; statistical analysis; text analysis; video retrieval; Chinese CCTV news video; Gray-based news video text extraction approach; OCR software package; mathematical statistical method; semantics understanding; text recognition; video analysis; video character extraction method; video clips retrieval; Character recognition; Data mining; Feature extraction; Image edge detection; Image segmentation; Optical character recognition software; Semantics; binaryzation; edge detection; projection; text extraction;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer Sciences and Convergence Information Technology (ICCIT), 2010 5th International Conference on
  • Conference_Location
    Seoul
  • Print_ISBN
    978-1-4244-8567-3
  • Electronic_ISBN
    978-89-88678-30-5
  • Type

    conf

  • DOI
    10.1109/ICCIT.2010.5711058
  • Filename
    5711058