基于小波分层连通树结构的信号重构

基于小波树模型的压缩感知可以通过较少的观测量得到鲁棒的信号重构，但采用最优树逼近时存在复杂度大的问题。在证明分层后的小波树仍然具备连通树性质的基础上，提出了基于小波分层连通树结构的压缩重构算法，在与原观测量一致的情况下，保证了重构精度并且提高了重构效率。实验结果表明，改进算法相对于原算法在处理大尺度数据时，效率有明显改善。

Compressive signal reconstruction using a hierarchical wavelet connected tree

The model-based compressive sensing (CS) dictates that robust signal recovery is possible from fewer measurements, but the computational complexity of this approach is large while using the optimal tree approximation with wavelets. In this paper, based on proving the property that the wavelet hierarchical tree still is connected, the model-based wavelet hierarchical connected tree CS algorithm is proposed. The proposed algorithm which has the equivalent measurements with that of model-based CS can enhance the signal-recovery efficiency and guarantee the signal-recovery accuracy. Numerical simulations demonstrate the validity of our new algorithm. Furthermore, the proposed algorithm has a distinct advantage when dealing with the mass of data.