Integrate Depth and Gray-level Information of Multi-View Video to Enhance Side Information of Distributed Video Coder

• Main Authors: Chi-chun Lu, 盧其均
• Other Authors: Jiann-jone Chen
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/r7rfgs

碩士 === 國立臺灣科技大學 === 電機工程系 === 99 === With the advance of video communication technology, the multimedia platform can not only receives and plays video streaming but also provides depth and stereo information of the natural scene for better visual perception. The 3D technologies play an important role of video innovation. The multi-view video coding (MVC) is an application of 3D video coder. However, the information amount of video data and the required computations for a multi-view system will be very large, as compared to single view videos. The distributed video coding (DVC) efficiently shifts computations to the decoder. The DVC decoder not only exploits inter- and intra view correlations but also utilizes depth information to enhance the quality of reconstructed images. We propose multi-view distributed video plus depth coding (MDVDC) to improve MDVC codec performance: (1) It applies depth and color perspective transform mapping algorithm to joint decoder and well utilizes temporal, interview and depth correlations to yield better side information (SI) images. (2) The proposed MDVDC algorithm can be applied to different GOPs and improve codec performance. It saves turbo decoding time. In comparisons, the Depth Homo-fusion is 0.3~3dB improved as compare to other methods when GOP=1. When GOP=2, the proposed Depth Mapping SIFT-BMP improved 2.3~8.35dB and 0.1~0.2dB as compare to MCTI and SIFT-BMP, respectively. It saves 20.05% decoding time as compares to MCTI.