• DocumentCode
    2918353
  • Title

    Spatial-domain unidirectional DVC with side-information dependent correlation channel estimation

  • Author

    Deligiannis, Nikos ; Munteanu, Adrian ; Clerckx, Tom ; Cornelis, Jan ; Schelkens, Peter

  • Author_Institution
    Dept. of Electron. & Inf. (ETRO), Vrije Univ. Brussel (VUB), Brussels, Belgium
  • fYear
    2009
  • fDate
    5-7 July 2009
  • Firstpage
    1
  • Lastpage
    6
  • Abstract
    Present distributed video coding (DVC) approaches assume the correlation channel to be side-information independent (SII). We have recently demonstrated that in reality the correlation channel in video is side-information dependent (SID) and that DVC systems which make SII assumptions suffer a significant performance penalty, depending entirely on the correlation statistics of the video data. In this paper, we present a novel spatial domain unidirectional DVC codec which is based on a novel SID technique enabling advanced correlation channel estimation at the decoder starting from a partial knowledge of it. Experimental results show that the proposed codec achieves superior performance compared to the state-of-the-art in Wyner-Ziv video coding while operating without a feedback channel.
  • Keywords
    channel estimation; correlation methods; statistical analysis; video coding; Wyner-Ziv video coding; advanced correlation channel estimation; correlation statistics; side-information dependent correlation channel estimation; spatial-domain unidirectional distributed video coding; video data; Channel estimation; Codecs; Computer architecture; Decoding; Informatics; Motion estimation; Source coding; State feedback; Statistical distributions; Video coding; Distributed video coding; correlation channel; overlapped block motion estimation; probabilistic reconstruction;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Digital Signal Processing, 2009 16th International Conference on
  • Conference_Location
    Santorini-Hellas
  • Print_ISBN
    978-1-4244-3297-4
  • Electronic_ISBN
    978-1-4244-3298-1
  • Type

    conf

  • DOI
    10.1109/ICDSP.2009.5201174
  • Filename
    5201174