Enhanced Recursive Residual Network for Single Image Super-Resolution

Single image super-resolution (SISR) based on deep learning methods has achieved great advance. Despite the great performance of these models, it is challenging to be applied to practical applications because of enormous parameters. In this paper, we propose an enhanced recursive residual network (ERRN) to address this problem. Specifically, based on residual networks, group convolution and recursive learning are adapted to reduce parameters. The results of evaluation on benchmark datasets show that the performance of ERRN is comparable with the state-of-art methods with much fewer parameters.