• DocumentCode
    527897
  • Title

    Application of partial reconfiguration of FPGAs in image processing

  • Author

    Raikovich, Tamás ; Fehér, Béla

  • Author_Institution
    Dept. of Meas. & Inf. Syst., Budapest Univ. of Technol. & Econ., Budapest, Hungary
  • fYear
    2010
  • fDate
    18-21 July 2010
  • Firstpage
    1
  • Lastpage
    4
  • Abstract
    FPGA based hardware accelerators have been more and more widely used in different kind of applications. As compared to other solutions and the direct hardware implementation, the advantage of the FPGA devices is their flexibility that arises from their programmable nature. In addition to this, some FPGA devices also support partial dynamic reconfiguration. When general purpose processors and reconfigurable computational structures are used together, the most advantageous properties of both components can be utilized. Algorithms (searching, sorting, signal and image processing, etc.) that are implemented such way have achieved 100 times or 1000 times better performance. This paper introduces a Xilinx Virtex-5 FPGA based reconfigurable hardware accelerator system, which has been created mainly for image processing purposes. By using the partial reconfiguration capability of the FPGA, the execution units in the system can be quickly reconfigured.
  • Keywords
    field programmable gate arrays; image processing; reconfigurable architectures; FPGA devices; Xilinx Virtex-5 FPGA; direct hardware implementation; execution units; field programmable gate array; general purpose processors; hardware accelerators; image processing; partial dynamic reconfiguration; reconfigurable computational structures; reconfigurable hardware accelerator system; Field programmable gate arrays; Finite impulse response filter; Hardware; Image processing; Pipelines; Pixel; FPGA; hardware accelerator; image processing; partial reconfiguration;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Ph.D. Research in Microelectronics and Electronics (PRIME), 2010 Conference on
  • Conference_Location
    Berlin
  • Print_ISBN
    978-1-4244-7905-4
  • Type

    conf

  • Filename
    5587101