基于3D-SPIHT的立体视频图像压缩编码

该文提出一种新的立体视频编码方案:在辅助序列中进行视差补偿预测和三维等级数集合分区(3D-SPIHT)编码,3D-SPIHT算法建立在真三维小波分解基础上,通过定义一种新的时空方向树结构,实现了静止图像SPIHT算法的三维扩展,实验结果表明该方案的编码性能略高于传统方案,具有较低的计算复杂度,所产生的嵌入式辅助序列码流,可根据通道带宽自适应调整输出码率,最大限度地提高辅助序列的质量。

An efficient coding algorithm for stereo video with 3D-SPIHT

This paper presents a new strategy for efficiently coding stereo video sequences. In the scheme, the auxiliary sequence is coded by disparity compensated prediction and 3D-SPIHT algorithm. 3D-SPIHT algorithm is based on genuine 3D wavelet transform, and implements a three dimensional extension of SPIHT in still image coding by defining a new spatio-temporal orientation tree structure. The scheme can provide comparable performance to the conventional method with minimal computational complexity and produce an embedded bit stream in the auxiliary sequence. According to channel bandwidth, the bit rate can be adaptively adjusted so that the best quality of the auxiliary sequence is obtained.