• DocumentCode
    2558498
  • Title

    Color pattern recognition using multiple phase-shifted reference joint transform correlation

  • Author

    Islam, M. Nazrul ; Karim, Mohammad A.

  • Author_Institution
    State Univ. of New York, Farmingdale, NY, USA
  • fYear
    2010
  • fDate
    7-7 May 2010
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    Optical joint transform correlation (JTC) technique has been proposed to be an efficient pattern recognition system. Color pattern recognition techniques utilize color of the target objects as additional feature to the spatial characteristics. A new optical JTC technique is developed for color pattern recognition employing the same reference image but with four different amounts of phase shift. A fringe-adjusted filter is also designed to enhance the discrimination between target and non-target objects. The proposed technique yields a very distinctive correlation output for color target detection. The simple architecture can easily be implemented using optical devices, which will provide a real-time pattern recognition decision. Performance of the proposed technique is investigated using computer simulation involving binary as well as gray images.
  • Keywords
    correlation methods; filtering theory; image colour analysis; object detection; transforms; color pattern recognition techniques; color target detection; computer simulation; fringe-adjusted filter; multiple phase-shifted reference joint transform correlation; optical devices; optical joint transform correlation technique; reference image; Color; Computer architecture; Computer simulation; Layout; Matched filters; Object detection; Optical devices; Optical filters; Pattern recognition; USA Councils; color pattern recognition; correlation; joint power spectrum; joint transform correlation;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Applications and Technology Conference (LISAT), 2010 Long Island Systems
  • Conference_Location
    Farmingdale, NY
  • Print_ISBN
    978-1-4244-5548-5
  • Electronic_ISBN
    978-1-4244-5550-8
  • Type

    conf

  • DOI
    10.1109/LISAT.2010.5478282
  • Filename
    5478282