基于多观测矢量和LDPCA的分布式视频编码

Distributed Video Coding Using Multiple Measurement Vectors and LDPCA

分布式视频编码与压缩感知技术都有助于将计算复杂度从编码端向解码端迁移,已有研究显示两者的结合可能提高分布式视频编码系统的性能。基于传统分布式视频编码框架,借助多观测矢量稀疏重建技术,设计了一种混合压缩感知和LDPCA的分布式视频编码系统。首先根据稀疏度对频带数据进行划分,再分模式进行编码表示,其中非零系数较多的低频频带使用LDPCA技术进行压缩;而对稀疏度较高的高频频带,使用多观测矢量的压缩感知技术以求进一步利用频带间的相关性。仿真结果表明,系统的性能优于LDPCA的变换域分布式视频编码系统。

Distributed video coding （DVC） is a new paradigm in video coding which shifts the computational burden from encoder to decoder. With the similar principle, Compressive Sensing （CS） was suggested to combine with DVC for achieving higher performance. However, the redundancy of some frequency bins is not sparse enough for CS codec to compress efficiently. On the other hand, the correlation of sparsity in different frequency bins can be consid- ered to further reduce the bit rate. In this paper, a hybrid Low-Density Parity-Check Accumulate code （LDPCA） and multiple measurement vectors （MMV） sparse recovery codec was proposed to deal with the respective sparsities. It was demonstrated that the MMV coding is an outstanding approach to make use of the correlation of sparsity. Simu- lation results verify that the performance of the proposed scheme outperforms that of the conventional LDPCA based scheme~