• DocumentCode
    2325426
  • Title

    Fast Mode Decision Algorithm for H.264/SVC Based on Motion Vector Relation Analysis

  • Author

    Sun, Xuesong ; Xiao, Song ; Du, Jianchao ; Hu, Min

  • Author_Institution
    ISN Nat. Key Lab., Xidian Univ., Xi´´an, China
  • fYear
    2011
  • fDate
    14-16 Oct. 2011
  • Firstpage
    278
  • Lastpage
    281
  • Abstract
    Recently, scalable video coding technique has attracted a lot of attention for its desired scalable functionality and coding efficiency. H.264/SVC encodes a video in a hierarchical way by which the coded bit stream can satisfy terminals with different capability of display and different condition of network at the same time. However, the huge computational complexity of mode decision causes difficulties in real-time coding. This paper proposed a fast mode decision algorithm for H.264/SVC by analyzing the relation of motion vectors of 8×8 blocks. The new algorithm leads an early termination and a prediction of Inter16×8 and Inter8×16. Experimental results show that the proposed algorithm can achieve a reduction of computational complexity by 63.28% on average, with the negligible 0.58% increase of Bit rate and 0.05 dB loss of PSNR.
  • Keywords
    computational complexity; data compression; decision theory; image motion analysis; video coding; H.264-SVC video encoding; PSNR loss; bit rate; computational complexity; fast mode decision algorithm; loss 0.05 dB; motion vector relation analysis; real-time coding; scalable video coding technique; Algorithm design and analysis; Bit rate; Encoding; PSNR; Prediction algorithms; Static VAr compensators; Vectors; Early Termination; H.264/SVC; Mode Decision; Motion Vector Relation;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP), 2011 Seventh International Conference on
  • Conference_Location
    Dalian
  • Print_ISBN
    978-1-4577-1397-2
  • Type

    conf

  • DOI
    10.1109/IIHMSP.2011.97
  • Filename
    6079582