• DocumentCode
    480941
  • Title

    A partial protection scheme based on layer dependency of Scalable Video Coding

  • Author

    Hendry ; Kim, Munchurl ; Sangjin Hahm ; Lee, Keunsik ; Park, Keunsoo

  • Author_Institution
    Information and Communications University (ICU), Daejeon, Republic of Korea
  • fYear
    2008
  • fDate
    July 29 2008-Aug. 1 2008
  • Firstpage
    777
  • Lastpage
    782
  • Abstract
    MPEG Scalable Video Coding (SVC), which also known as H.264 Scalable Extension, makes video adaptation more flexible to meet user consumption preferences, application requirements, and also varying network conditions. An SVC bit stream consists of one or more scalable layers that form hierarchical dependency such that a layer has dependency on its lower layers. Therefore a layer cannot be decoded perfectly if all of its lower layers are not decoded a priori. In this paper, we propose layered protection, which is a partial protection scheme, for SVC bit streams by fully utilizing layer dependency characteristics of SVC. Layered protection scheme can achieve protection degree as high as that of given by protection of entire SVC bit streams (referred as total protection). Therefore it reduces complexity for protecting/unprotecting SVC bit streams. We use a modified Structural Similarity Index (SSIM) to show effectiveness of layered protection scheme in term of visual perception degradation. Our experiments show that layered protection scheme is not only effective in term of visual degradation, but also efficient in term of computational complexity. For our test videos, in average, we only need to protect 42.45% of input bit stream to achieve the same visual perception degradation as that of given by total protection (protecting 100% of input bit stream).
  • Keywords
    H.264 Scalable Extension; Layered Protection; MPEG SVC; Partial Protection; Video Protection;
  • fLanguage
    English
  • Publisher
    iet
  • Conference_Titel
    Visual Information Engineering, 2008. VIE 2008. 5th International Conference on
  • Conference_Location
    Xian China
  • ISSN
    0537-9989
  • Print_ISBN
    978-0-86341-914-0
  • Type

    conf

  • Filename
    4743524