• DocumentCode
    2836675
  • Title

    High performance H.264/AVC encoding motion prediction algorithm

  • Author

    Husemann, Ronaldo ; Roesler, Valter ; Kintschner, Ricardo ; Fröhlich, Humberto ; Susin, Altamiro A.

  • Author_Institution
    Inst. of Inf. II, UFRGS, Porto Alegre, Brazil
  • fYear
    2011
  • fDate
    11-14 Sept. 2011
  • Firstpage
    957
  • Lastpage
    960
  • Abstract
    The ITU-T H.264 can be considered currently as the most efficient professional video coding standard. This codec combines distinct complex techniques in order to encapsulate raw video into efficient compressed streams. Unfortunately, the improved H.264 coding efficiency increases significantly the computational complexity, making real-time applications difficult to be achieved. Particularly the motion prediction is an important and complex algorithm used to remove temporal redundancy of image sequences. Considering that, we propose an optimized high performance small diamond topology zonal search motion prediction algorithm. The proposed method has been designed to operate with orthogonal linear vectors, which are specially aligned with 4:1 sub sampled memories organization reducing the number of operations and data manipulations. The proposed solution was implemented in the H.264 JM reference software obtaining results at least two times faster than other available fast algorithms with negligible quality losses.
  • Keywords
    computational complexity; video coding; H.264 JM reference; ITU-T H.264; complex algorithm; computational complexity; data manipulations; h.264/AVC encoding algorithm; motion prediction; orthogonal linear vectors; video coding; Algorithm design and analysis; Arrays; Diamond-like carbon; Image coding; Motion estimation; Prediction algorithms; Vectors; H264/AVC; Motion compensation; video encoding;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Image Processing (ICIP), 2011 18th IEEE International Conference on
  • Conference_Location
    Brussels
  • ISSN
    1522-4880
  • Print_ISBN
    978-1-4577-1304-0
  • Electronic_ISBN
    1522-4880
  • Type

    conf

  • DOI
    10.1109/ICIP.2011.6116720
  • Filename
    6116720