• DocumentCode
    2933359
  • Title

    An Improved Spatial Error Concealment Algorithm Based on H.264

  • Author

    Jin-wang, Yi ; En, Cheng ; Fei, Yuan

  • Author_Institution
    Key Lab. of Underwater Acoust. Commun. & Marine Inf. Technol., Xiamen Univ., Xiamen, China
  • Volume
    3
  • fYear
    2009
  • fDate
    21-22 Nov. 2009
  • Firstpage
    455
  • Lastpage
    458
  • Abstract
    The losses of packets are inevitable when the video is transported over error-prone networks. Error concealment methods can reduce the quality degradation of the received video by masking the effects of such errors. This paper presents a novel spatial error concealment algorithm based on the directional entropy in the available neighboring Macro Blocks (MBs), which can adaptively switch between weighted pixel average (WPA) adopted in H.264 and an improved directional interpolation algorithm to recover the lost MBs. In this work, the proposed algorithm was evaluated on H.264 reference software JM8.6. The illustrative examples demonstrate that the proposed method can achieve better Peak to Signal-to-Noise Ratio (PSNR) performance and visual quality, compared with WPA and the conventional directional interpolation algorithm respectively.
  • Keywords
    entropy; interpolation; video coding; H.264; directional entropy; directional interpolation; error-prone networks; macro blocks; peak to signal-to-noise ratio; spatial error concealment; weighted pixel average; Computer errors; Degradation; Entropy; IEC standards; ISO standards; Image reconstruction; Information technology; Interpolation; Switches; Transform coding; directional entropy; directional interpolation; image recovery; sptial error concealment; weighted pixel average;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Information Technology Application, 2009. IITA 2009. Third International Symposium on
  • Conference_Location
    Nanchang
  • Print_ISBN
    978-0-7695-3859-4
  • Type

    conf

  • DOI
    10.1109/IITA.2009.451
  • Filename
    5370364