• DocumentCode
    2042779
  • Title

    Flast identification of player position in soccer broadcast video by block-based camera view angle search

  • Author

    Mochizuki, Takahiro ; Fujii, Mahito ; Shibata, Masahiro ; Sakai, Yoshinori

  • Author_Institution
    NHK Sci. & Tech. Res. Labs., Tokyo, Japan
  • fYear
    2009
  • fDate
    16-18 Sept. 2009
  • Firstpage
    408
  • Lastpage
    413
  • Abstract
    We propose a way of quickly and automatically identifying the positions of players in soccer broadcast video. The approach is based on fast specification of camera view angles corresponding to frame pictures by using a ldquoblock-based searchrdquo. Our method can decide on a camera view angle that corresponds to a frame picture by making block-based connections for how to allocate ldquoline-blocksrdquo, i.e., block images on field lines, between the frame picture and the soccer field space. We can obtain the position of the players by projecting the ldquoplayer-blockrdquo positions to the soccer field space along the appropriate camera view angle specified by the search. Our method can be used to make these processes faster than conventional ones that are based on computationally expensive processing such as line extraction by Hough transform. The proposal is evaluated from the aspects of both accuracy and computational cost by experiments using soccer broadcast videos.
  • Keywords
    Hough transforms; video signal processing; video streaming; Hough transform; block-based camera view angle search; block-based connection; frame pictures; line extraction; player position identification; player-block position; soccer broadcast video; soccer field space; Broadcast technology; Cameras; Computational efficiency; Costs; Games; Laboratories; Layout; Multimedia communication; Proposals; TV broadcasting;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Image and Signal Processing and Analysis, 2009. ISPA 2009. Proceedings of 6th International Symposium on
  • Conference_Location
    Salzburg
  • ISSN
    1845-5921
  • Print_ISBN
    978-953-184-135-1
  • Type

    conf

  • DOI
    10.1109/ISPA.2009.5297705
  • Filename
    5297705