Abstract: A multiple description scalable video coding based on directional lifting wavelet transform is proposed, it can enhance the error-resilient and network-adaptive ability of video coding. It divides each frame into two subframes to form two subsequences by using quincunx down-sampling, each subsequence is coded by a scalable 3D wavelet video coder to form one description. To improve coding efficiencies of the subframes, directional lifting wavelet transform is used to perform spatial decomposition. A five mode directional lifting is introduced to improve coding efficiency. Experimental results prove the validity of the proposed algorithm.

Key words: Multiple Description Scalable Coding(MDSC), quincunx down-sampling, 3D wavelet video coding, plane prediction mode, image interpolation

摘要： 提出一种基于方向提升小波变换的多描述可分级视频编码方法，该方法能增强视频编码的容错性和网络自适应性。对视频序列中的每一帧图像按梅花下采样产生2个子序列，对每个子序列分别进行三维小波可分级视频编码形成2个描述。为解决传统二维小波变换方法对梅花下采样图像编码效率低的问题，引入方向提升小波变换对图像进行空域分解，提出5模式方向提升小波变换以进一步提高编码效率。实验结果证明了该方法的有效性。

关键词: 多描述可分级编码, 梅花下采样, 三维小波视频编码, 平面预测模式, 图像插值

CLC Number: 

• TN919.8