• DocumentCode
    592060
  • Title

    Boundary Matching Based Spatial Interpolation for Consecutive Block Loss Concealment

  • Author

    Shaoshuai Gao

  • Author_Institution
    Sch. of Electron. & Commun. Eng., Univ. of Chinese Acad. of Sci., Beijing, China
  • fYear
    2012
  • fDate
    10-12 Dec. 2012
  • Firstpage
    128
  • Lastpage
    132
  • Abstract
    This paper presents a spatial error concealment algorithm for recovering the lost packets in still images or Intra-coded frames in video. The idea is based on the work proposed by Hsia [1], where a 1-D boundary matching is applied to recover the lost areas while preserving the edge information. However, we modify Hsia´s algorithm in several aspects to improve its performance in terms of subjective and objective quality, i.e., smoothing operation by an averaging filter, adaptive searching range expansion, and matching direction crossing removing. Simulation results shows superiority of the proposed algorithm compared with Hsia´s algorithm and some other state-of-the-art algorithms.
  • Keywords
    edge detection; image matching; interpolation; smoothing methods; video coding; 1D boundary matching; Hsia algorithm; adaptive searching range expansion; averaging filter; boundary matching-based spatial interpolation; consecutive block loss spatial error concealment algorithm; edge information preservation; intracoded video frames; lost packet recovery; matching direction crossing removal; objective quality; performance improvement; smoothing operation; still images; subjective quality; Complexity theory; Image edge detection; Image reconstruction; Interpolation; PSNR; Signal processing algorithms; Streaming media; block loss recovery; robust video communication; spatial error concealment;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Multimedia (ISM), 2012 IEEE International Symposium on
  • Conference_Location
    Irvine, CA
  • Print_ISBN
    978-1-4673-4370-1
  • Type

    conf

  • DOI
    10.1109/ISM.2012.32
  • Filename
    6424645