The Compression Sensing Reconstruction for the 1-d Signal Based on Non-local 1-d Convolution Applied to LISTA
Keywords:
Compression sensing reconstruction;, 1-d signal, LISTA, non-local 1-d convolutionAbstract
The compression sensing reconstruction for the 1-d signal can contribute to the communication of autonomous driving, intelligent robots, and fire exploration robots. This paper designs a compression sensing reconstruction method for the 1-d signal. To reconstruct 1-d signal, the deep learning method LISTA is used. Then, the linear convolution layer in LISTA is improved by non-local 1-d convolution. The convolution regions of improved non-local 1-d convolution contain: 1) the convolution before; 2) the current convolution; and 3) the convolution after. Experimental results show the reconstructed results of LISTA is better than those of ISTA and the reconstructed results of LISTA-n1d is better than those of LISTA and the reconstruction time of ISTA, LISTA, and LISTA-n1d are 0.1s, 0.12s, and 0.17s, respectively.



