基于结构化稀疏模型的压缩感知重构改进算法

New recovery algorithm for compressed sensing based on structured sparse model

目前,标准的CS重构算法仅利用信号和图像在小波变换下的稀疏先验信息,而并没有利用变换系数具有的结构化特性。为了能够快速精确地重建原始信号,将结构化稀疏模型与SP算法、CoSaMP算法相结合,提出了压缩感知重构的改进算法。另外,将基于双树复小波变换的系数结构模型融入上述算法,进一步提高重构性能。实验结果表明,所提出的算法可获得更高的图像重建质量。

Recently, normal recovery algorithms for CS only use signal and image sparse priors under wavelet, make no use of the tree structure priors. In order to reconstruct the original signal quickly and accurately, this paper brings the tree structure sparse model into SP algorithm, CoSaMP-algorithm and gets the improved recovery algorithm for compressed sensing. Combin- ing with structured sparse model and dual-tree complex wavelet transform, a new recovery algorithm for CS is proposed. The simulated results show that the algorithm can achieve higher reconstructed image performance.