基于双边信息的分布式视频压缩感知模型研究

Study On DistributedSCompressed Video SensingS Based On Double Side Information

摘要:

针对传统视频编码技术计算量大和复杂度高的缺点,提出一种基于双边信息的分布式视频压缩感知算法。该算法将压缩感知技术与分布式视频编码技术相结合,把视频序列分为Key帧和CS帧,Key帧运用传统的帧内编码和解码,CS帧编码端运用压缩感知编码,解码端运用视频块内与视频块间的双边信息和梯度投影算法进行优化重构。通过双边信息的运动估计和压缩编码器的设计,实现基于双边信息的分布式视频压缩感知模型的构建。仿真结果表明该模型既可以实现高效编码,又可以实现复杂度由编码端向解码端转移,在较低的采样率下,提高视频的压缩能力和传输速度。

Abstract:

In view of the shortcomings of the traditional video coding technology of large amount of calculation and high complexity, this paper presented based on bilateral distributed video compressed sensing algorithm. The algorithm combined compressed sensing technology and distributed video coding technology. The video sequence was divided into Key frames and CS frames. Key frames used traditional intra coding and decoding, and CS frames coding end used compressed sensing encoding, decoding end used intra-view and Inter-view of bilateral information and gradient projection to optimize and reconstruct information. Distributed compressive video sensing model that based on bilateral information was builded by bilateral motion estimation information and compression encoder design. The simulation results showed that this model can achieve high-efficiency encoding, while the implementation complexity of the decoding by the encoding end to end of the transfer, at a lower sampling rate, the ability to improve video compression and transmission speed.