Image Super-Resolution Reconstruction Algorithm Based on FSRCNN and Residual Network

Abstract

In order to effectively improve the resolution of the image and restore more edge and texture details, an image super-resolution reconstruction algorithm based on FSRCNN and residual network is proposed. Firstly, a deep channel is constructed, based on the FSRCNN network, a simple residual network is added to estimate the high-frequency information, and the output of FSRCNN network is used as the low-frequency information to be combined with the high-frequency information to obtain a reconstructed image. Then, the information distillation module is fused as a shallow channel to further estimate the edge and texture details. Finally, the outputs of the two channels are combined to obtain the final reconstructed image. Experiments results illustrate that, compared with the FSRCNN approach, the proposed approach has better subjective visual effects corresponding to the reconstruct images on the Set5, Set14 and B100 data sets, and its average PSNR is increased by 0.26dB, 0.18dB and 0.15dB, respectively.