Image Reconstruction from Patch Compressive Sensing Measurements

Yahan Wang, Huihui Bai, Yao Zhao

Abstract: The compressive sensing theory has been successfully applied to image compression in the past few years. Recently, deep network-based compressive sensing image reconstruction algorithms have been proposed, which reduce the computational complexity compared with traditional iterative reconstruction algorithms. But most of those are patch-based reconstruction methods, which leads to blocky artifacts for the full image assembled by patch reconstruction. In this paper, we propose a novel image reconstruction network (CSReNet) from patch compressive sensing measurements. Different from other deep network-based algorithms, our network can not only recovery image from patch compressive sensing measurements, also remove the blocky artifacts. There are two modules, reconstruction module and removal module in our network. Experimental results on test data show that our proposed network outperforms several compressive sensing reconstruction algorithms with patch-based CS measurements.