Image Super-Resolution via Deep Recursive Residual Network

TLDR

Convolutional neural networks have achieved great success in single image super‑resolution by learning a nonlinear mapping from low‑resolution to high‑resolution images, but this typically requires a large number of parameters. This work introduces the Deep Recursive Residual Network (DRRN), a very deep CNN with up to 52 layers designed to be both deep and parameter‑efficient. DRRN employs global and local residual learning to ease training of deep networks, and uses recursive learning to increase depth while keeping the parameter count low. Benchmark tests show that DRRN significantly outperforms state‑of‑the‑art SISR methods while using far fewer parameters. Code is available at https://github.com/tyshiwo/DRRN_CVPR17.

Abstract

Recently, Convolutional Neural Network (CNN) based models have achieved great success in Single Image Super-Resolution (SISR). Owing to the strength of deep networks, these CNN models learn an effective nonlinear mapping from the low-resolution input image to the high-resolution target image, at the cost of requiring enormous parameters. This paper proposes a very deep CNN model (up to 52 convolutional layers) named Deep Recursive Residual Network (DRRN) that strives for deep yet concise networks. Specifically, residual learning is adopted, both in global and local manners, to mitigate the difficulty of training very deep networks, recursive learning is used to control the model parameters while increasing the depth. Extensive benchmark evaluation shows that DRRN significantly outperforms state of the art in SISR, while utilizing far fewer parameters. Code is available at https://github.com/tyshiwo/DRRN_CVPR17.