自适应PCA稀疏基底的分布式视频压缩感知重构

Distributed Video Compressive Sensing Reconstruction Based on Adaptive PCA Sparse Basis

为了提高分布式视频压缩感知(Distributed Video Compressive Sensing,DVCS)的率失真性能,提出利用自适应稀疏基底进行联合重构。算法利用帧间运动信息形成样本数据矩阵,再利用主成分分析(Principle Components Analysis,PCA)训练出其显著主成分构成稀疏字典,该稀疏字典不仅可根据视频时空统计特征自适应变化而且可有效地抑制噪声。仿真实验表明,该联合重构算法可有效地改善主客观视频重构质量,能够以一定的计算复杂度为代价提高DVCS系统的率失真性能。

To improve the rate-distortion performance of distributed video compressive sensing （DVCS）, the adaptive sparse basis is proposed to reconstruct the video signal in this paper. The proposed algorithm firstly generates the example-based data matrix using the motion information between frames, and then uses principle components analysis （PCA） to train its some significant principle components from which the sparse dictionary is constructed. This sparse dictionary can not only adaptively change in terms of the spatial-temporal characteristics, but also has ability to suppress noises. The simulation experiments show that the proposed joint reconstruction algorithm can effectively improve the objective and subjective quality of video, and enhance the rate-distortion performance of DVCS system at the cost of a certain computational complexity.