基于变采样率压缩感知的视频压缩研究  

Block compressed sensing of video based on variable sampling rates

作　　者：李如春[1] 李林[1] 常丽萍[1] 

机构地区：[1]浙江工业大学信息工程学院,浙江杭州310023

出　　处：《电子技术应用》2015年第10期147-149,153,共4页Application of Electronic Technique

基　　金：2013年青年科学基金项目(61304124)

摘　　要：业界分块视频压缩感知通常对所有图像块均采用相同的测量矩阵进行测量,这种方式未考虑到视频中不同区域的变化程度不同的事实。在视频帧间相关性的基础上提出一种自适应分配采样率的方法,即在编码端根据图像块的帧间相关性大小分类并分配不同的采样率;在解码端使用全变差算法以充分利用帧间相关性。为减小网络环境影响,此算法不区分参考帧与非参考帧,并对每一帧作相同处理。实验结果表明,该方法能够在较低采样率下重构出较高质量的视频图像,并且缩短计算时间。The current block compressed sensing of video usually uses the same measurement matrix to all image block, this method ignores the fact that the structural complexity and the movement varies from different regions in video. Therefore, an adaptive allocation of sampling rate compressed sensing method is proposed according to the distribution feature of the correlations between neighboring frames. It classified blocks into different types depending on the inter-frame correlation, and adjusted different sampling rate to different blocks, total variation algorithm was used to reconstruct the videos to make fully use of the inter-frame correlation. In order to overcome the network environment, this algorithm didn＇t distinguish the reference frame and the non-reference frame, each frame was treated equally. The experimental results show that the method can reconstruct high quality video image under low sampling rate, and with the variable sampling rate measurement method, a higher reconstruction quality can be achieved for the regions containing relatively fast movement.

关 键 词：压缩感知 视频编码 全变差算法 变采样率 分块 

分 类 号：TN911.73[电子电信—通信与信息系统]