• DocumentCode
    73224
  • Title

    Quality-Efficient Upsampling Method for Asymmetric Resolution Stereoscopic Video Coding With Interview Motion Compensation and Error Compensation

  • Author

    Kuo-Liang Chung ; Yong-Huai Huang ; Wen-Chang Liu

  • Author_Institution
    Dept. of Comput. Sci. & Inf. Eng., Nat. Taiwan Univ. of Sci. & Technol., Taipei, Taiwan
  • Volume
    24
  • Issue
    3
  • fYear
    2014
  • fDate
    Mar-14
  • Firstpage
    430
  • Lastpage
    442
  • Abstract
    In asymmetric resolution stereoscopic video coding (ARSVC), to reduce the bitrate required for bandwidth-limited channels, each downsampled right-view frame is a quarter the size of the corresponding left-view frame and will be upsampled to the original size by the decoder. In this paper, two upsampling methods for ARSVC are proposed. The first proposed method integrates the traditional Wiener filter-based method and the interview prediction scheme by incorporating the information from the similarity between the left-view and right-view frames. By compensating for the prediction errors, the second proposed method further improves the quality of the upsampled images obtained by the first proposed method, especially in the sequences with heavy irregular textures.
  • Keywords
    Wiener filters; image resolution; image texture; stereo image processing; video coding; visual perception; ARSVC; Wiener filter based method; asymmetric resolution stereoscopic video coding; bandwidth limited channels; error compensation; interview motion compensation; irregular textures; quality efficient upsampling method; Bit rate; Decoding; Encoding; Error compensation; Stereo image processing; Video coding; Video sequences; 6-tap filter; Wiener filter; asymmetric resolution stereoscopic video coding (ARSVC); bitrate reduction; error compensation; interview prediction; three-dimensional television (3-D-TV); upsampling;
  • fLanguage
    English
  • Journal_Title
    Circuits and Systems for Video Technology, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1051-8215
  • Type

    jour

  • DOI
    10.1109/TCSVT.2013.2276877
  • Filename
    6575155