运动补偿预测残差稀疏重构的压缩视频传感

针对低复杂度视频编码需求,基于压缩传感(Compressive Sensing,CS)理论,提出了一种分布式压缩视频传感算法。低复杂度的编码器独立随机投影关键帧和CS帧,采集压缩视频数据;在解码端进行运动补偿预测以利用帧间相关性,对预测残差稀疏重构实现CS帧重建。仿真测试表明,与现有的3种压缩视频传感算法相比,所提算法重建的视频质量更好,适合无线视频监控及无线视频传感网络等应用。

Compressive Video Sensing via Sparse Recovery of Motion Compensated Prediction Residual

In order to meet the requirement of low-complexity video encoding,a distributed compressive video sensing algorithm based on the compressive sensing(CS) theory is proposed.The algorithm directly captures compressed video data at a low-complexity encoder by randomly projection key frames and CS frames.Motion compensation prediction by exploiting inter-frame comelatims is performed and Reconstruction of CS frames by utilizing sparse recovery of prediction residual is implemented at the decoder.Simulation results show that the proposed algorithm outperforms three known compressive video sensing algorithms and is a promising candidate for wireless visual surveillance and visual sensor networks.