Abstract
With the help of deep convolutional neural networks, a vast majority of single image super-resolution (SISR) methods have been developed, and achieved promising performance. However, these methods suffer from over-smoothness in textured regions due to utilizing a single-resolution network to reconstruct both the low-frequency and high-frequency information simultaneously. To overcome this problem, we propose a Multi-resolution space-Attended Residual Dense Network (MARDN) to separate low-frequency and high-frequency information for reconstructing high-quality super-resolved images. Specifically, we start from a low-resolution sub-network, and add low-to-high resolution sub-networks step by step in several stages. These sub-networks with different depth and resolution are utilized to produce feature maps of different frequencies in parallel. For instance, the high-resolution sub-network with fewer stages is applied to local high-frequency textured information extraction, while the low-resolution one with more stages is devoted to generating global low-frequency information. Furthermore, the fusion block with channel-wise sub-network attention is proposed for adaptively fusing the feature maps from different sub-networks instead of applying concatenation and $1\times 1$ convolution. A series of ablation investigations and model analyses validate the effectiveness and efficiency of our MARDN. Extensive experiments on benchmark datasets demonstrate the superiority of the proposed MARDN against the state-of-the-art methods. Our super-resolution results and the source code can be downloaded from https://github.com/Periter/MARDN .
Highlights
Single image super-resolution (SISR) aims to estimate a highresolution counterpart from a single low-resolution image and has been researched for decades in both industry and academy
We propose a novel SISR method based on Multi-resolution space-Attended Residual Dense Network (MARDN) to extract different frequency representations in parallel for reconstructing the final super-resolved images
Our MARDN performs similar to RCAN and second-order attention network (SAN) and better than the other stateof-the-art convolutional neural networks (CNNs)-based methods
Summary
Single image super-resolution (SISR) aims to estimate a highresolution counterpart from a single low-resolution image and has been researched for decades in both industry and academy. SISR is a notoriously challenging ill-posed problem because a particular low-resolution image may correspond to a crop of plausible high-resolution counterparts. To resolve this undetermined problem, plenty of SISR methods e.g., [4]–[17] have been proposed, among which the Convolutional Neural Network. RDN [13] uses residual dense blocks to extract hierarchical feature maps. CNF [18] constructs different depth parallel context-wise network fusion framework. These methods suffer from varying degrees of over-smoothness in textured regions due to utilizing a single-resolution network to reconstruct both the low-frequency and high-frequency information.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have