• DocumentCode
    3179255
  • Title

    H.264 deblocking filter enhancement

  • Author

    Mahjoub, Waseim Hashem ; Osman, Hossam ; Aly, Gamal M.

  • Author_Institution
    Software Eng. Competence Center, Inf. Technol. Ind. Dev. Agency, Giza, Egypt
  • fYear
    2011
  • fDate
    Nov. 29 2011-Dec. 1 2011
  • Firstpage
    219
  • Lastpage
    224
  • Abstract
    This paper proposes new software-based techniques for speeding and reducing the complexity of the deblocking filter used in the state-of-the-art H.264 international video coding standard to improve the visual quality of the decoded video frames. The proposed techniques are classified as standard-compliant and standard-noncompliant techniques. The standard-compliant techniques optimize the standard filter through optimizing the boundary strength calculation and group filtering of macroblocks. The standard-noncompliant techniques predict the new boundary strength and edge detection conditions from previous values. Experimental results on both an embedded platform and a desktop PC show significant increment in performance improvement that reaches 47% for the standard-compliant techniques and 80% for the standard-noncompliant techniques. They also demonstrate that for standard-noncompliant techniques the quality degradation computed using the Peak Signal to Noise Ratio is insignificant.
  • Keywords
    data compression; edge detection; filtering theory; video coding; H.264 deblocking filter enhancement; H.264 international video coding standard; deblocking filter; decoded video frames; desktop PC; edge detection; peak signal to noise ratio; software-based techniques; standard-noncompliant techniques; visual quality; Equations; Filtering; Mathematical model; Transform coding; Vectors; Deblocking filter; H.264; Optimization; Performance improvement; Video coding;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computer Engineering & Systems (ICCES), 2011 International Conference on
  • Conference_Location
    Cairo
  • Print_ISBN
    978-1-4577-0127-6
  • Type

    conf

  • DOI
    10.1109/ICCES.2011.6141046
  • Filename
    6141046