分布式压缩感知视频编码中CS帧的二次修正准则研究

The Study of CS Frames' s Second Correlation Rule of Distributed Video Compression Coding

分布式视频编码是新的视频编码体系,与传统的视频编码体系相比,具有编码端相对简单、解码端相对复杂的特点。此外,压缩感知突破了奈奎斯特采样定理,降低了信号的采样率。将压缩感知理论应与分布式视频编码相结合,使编码端复杂度降低。在一些分布式压缩视频编码研究中,CS帧是由边信息和发送端传送的信息联合重建的,由于不同CS帧的边信息的预测准确度不同,导致不同CS帧恢复质量相差较大。为了解决这个问题,本文对CS帧的二次修正准则的进行研究,首先从理论上推导出方差作为修正准则的可行性,并在实验中加以验证。实验可知,本文提出的方法在一定程度上改善了这些帧的重建质量。

Distributed video coding is a new video coding system. It has a simpler coding and more complex decoding compared with the traditional video coding system. In addition, the compressed sensing breaks the Nyquist sampling theorem reducing the sampling rate of the signal. Combining the compressed sensing theory and distributed video coding reduce the complexity of the coding. In some study of the distributed compressed video coding, the CS frames are reconstructed with the help of the side information. Due to the different accuracy estimation of the CS frames, the reconstructed CS frames have different quality. This paper studies the criteria of secondary correction of CS frames to solve this problem. Firstly, this paper study the feasibility of the variance criteria and proof it in experiences. The Experiment shows that the variance criteria improves the reconstructed quality of CS frames.