Feature Preserving and Enhancing Network for Image Super-Resolution

The success of modern single image super-resolution (SISR) algorithms is inspired by the development of deep convolutional neural networks (CNNs). However, these CNN-based methods require considerable computation and complexity, making it impossible for these methods to perform real-time calculations in edge devices. Thus, lightweight model design has become a development trend in the super-resolution field, including pruning, quantization, and other methods. The 1-bit quantization is an extreme lightweight method which can reduce the calculation amount of the model in an extreme manner and is friendly to hardware such as edge devices. Most existing binary quantization approaches lead to a large information loss during forward propagation, especially in detailed color information (e.g., edge, texture, and contrast). The loss of color information makes modern binary methods unsuitable for SISR tasks. We think the loss occurs because these methods typically utilize a uniform threshold to quantize the weights and activations. Thus, in this article, we thoroughly analyze the difference between normal classification tasks and SISR tasks, and present a binarization scheme based on local means. The proposed method can maintain more detailed information in feature maps using dynamic thresholds during quantization. Specifically, each value in the full precision activations has a corresponding threshold during the quantization process, and those thresholds are determined by the full precision values of the surroundings. In addition, a gradient approximator is introduced to adaptively optimize the gradient for updating binary weights. We then verify the effectiveness of our method for training binary networks on several SISR benchmarks including VDSR and SRResNet. Experimental results show that the proposed method can outperform the state-of-the-art algorithms to obtain binary networks for image super-resolution with better peak signal-to-noise ratio (PSNR) values and visual quality.

Single image super-resolution using a deep encoder–decoder symmetrical network with iterative back projection

In recent years, single image super-resolution (SISR) methods using deep convolution neural network (CNN) have achieved impressive results. Thanks to the powerful representation capabilities of the deep networks, numerous previous ways can learn the complex non-linear mapping between low-resolution (LR) image patches and their high-resolution (HR) versions. However, excessive convolutions will limit the application of super-resolution technology in low computing power devices. Besides, super-resolution of any arbitrary scale factor is a critical issue in practical applications, which has not been well solved in the previous approaches. To address these issues, we propose a lightweight information multi-distillation network (IMDN) by constructing the cascaded information multi-distillation blocks (IMDB), which contains distillation and selective fusion parts. Specifically, the distillation module extracts hierarchical features step-by-step, and fusion module aggregates them according to the importance of candidate features, which is evaluated by the proposed contrast-aware channel attention mechanism. To process real images with any sizes, we develop an adaptive cropping strategy (ACS) to super-resolve block-wise image patches using the same well-trained model. Extensive experiments suggest that the proposed method performs favorably against the state-of-the-art SR algorithms in term of visual quality, memory footprint, and inference time. Code is available at \url{https://github.com/Zheng222/IMDN}.

Lightweight single image super-resolution with attentive residual refinement network

The deep convolutional neural network has achieved great success in the Single Image Super-resolution task. It is obviously that among the well-known super-resolution methods, the deep learning-based algorithms show the most advanced performance. However, the most advanced algorithms currently use complex networks with a large number of parameters, which makes it difficult to apply deep learning algorithms on mobile devices. To solve this problem, we propose a lightweight dual-residual network(LDRN) for single image super-resolution, which has better reconstruction quality than most current advanced lightweight algorithms. Due to its fewer parameters and computational expense, real-time and mobile applications of our networks can be easily realized. On the basis of the residual module, we propose a new residual unit, which uses two depthwise separable(DW) convolution to obtain better balance between feature extraction capacity and lightweight performance. We further design a dual-stream residual block, which contains a multiplication branch and an addition branch. The dual-stream residual block can improve the reconstruction performance more effectively than expanding the network width. In addition, we also designed a new up-sampling module to simplify the previous up-sampling methods. Extensive experimental results show that our network has better reconstruction performance and lightweight performance than most existing state-of-the-art algorithms. Our code is available at https://github.com/Jiangyichun-cust/pytorch-LDRN.

Recently, convolutional neural network (CNN) based models have shown great potential in the task of single image super-resolution (SISR). However, many state-of-the-art SISR solutions are reproducing some tricks proven effective in other vision tasks, such as pursuing a deeper model. In this paper, we propose a new solution (named as Multi-Receptive-Field Network - MRFN), which outperforms existing SISR solutions in three different aspects. First, from receptive field: a novel multi-receptive-field (MRF) module is proposed to extract and fuse features in different receptive fields from local to global. Integrating these hierarchical features can generate better mappings on recovering high-fidelity details at different scales. Second, from network architectures: both dense skip connections and deep supervision are utilized to combine features from the current MRF module and preceding ones for training more representative features. Moreover, a deconvolution layer is embedded at the end of the network to avoid artificial priors induced by numerical data pre-processing (e.g., bicubic stretching), and speed up the restoration process. Finally, from error modeling: different from $L1$ and $L2$ loss functions, we proposed a novel two-parameter training loss called Weighted Huber loss function which can adaptively adjust the value of back-propagated derivative according to the residual value, thus fit the reconstruction error more effectively. Extensive qualitative and quantitative evaluation results on benchmark datasets demonstrate that our proposed MRFN can achieve more accurate recovering results than most state-of-the-art methods with significantly less complexity.

Traditional works have shown that patches in a natural image tend to redundantly recur many times inside the image, both within the same scale, as well as across different scales. Make full use of these multi-scale information can improve the image restoration performance. However, the current proposed deep learning based restoration methods do not take the multi-scale information into account. In this paper, we propose a dilated convolution based inception module to learn multi-scale information and design a deep network for single image super-resolution. Different dilated convolution learns different scale feature, then the inception module concatenates all these features to fuse multi-scale information. In order to increase the reception field of our network to catch more contextual information, we cascade multiple inception modules to constitute a deep network to conduct single image super-resolution. With the novel dilated convolution based inception module, the proposed end-to-end single image super-resolution network can take advantage of multi-scale information to improve image super-resolution performance. Experimental results show that our proposed method outperforms many state-of-the-art single image super-resolution methods.

Recently deep convolutional neural network (CNN) has made significant achievement in Single Image Super-Resolution (SISR). Most CNN-based SISR methods used the default L2 norm of the error. However, for Hyperspectral Image (HSI), this loss function may bring spectral inconsistencies. The main reason is that most methods did not pay much attention to spectral loss. To HSI, the loss function should capture not only spatial information but also spectral consistency. In this paper, a Multi-Losses Function Network (MLFN) simultaneously considering spatial and spectral information is proposed, and is composed of two parts: one is Concatenate Dense Residual Network (CDRN), and the other is Loss Network (LN). CDRN is an image reconstruction network which can utilize the hierarchical features extracted from the low-resolution image. LN includes pixel-wise spatial loss and spectral loss which drive the learning of the entire reconstruction model. The experimental results prove that the proposed MLFN can enhance spatial resolution with the consistency of the spectrum of HSI preserved.

Blind super-resolution (SR) of blurry and noisy low-resolution (LR) images is still a challenging problem in single image super-resolution (SISR). The performance of most existing convolutional neural network (CNN)-based models is inevitably degraded when LR images are corrupted by both blur and noise. For those blind SR methods based on kernel estimation, accurate estimation is barely attained under complex degradations and this gives rise to poor-quality results. To address these problems, we propose a deep progressive network under a probabilistic framework and a novel up-sampling method for blind super-resolution with multiple degradations, which effectively utilizes image priors across scales. Experimental results show that the proposed method achieves promising performance on images with multiple degradations.

In recent years, deep learning methods, especially deep convolutional neural network, have been successfully applied to the single‐image super‐resolution (SISR) task. In this Letter, the authors propose an accurate SISR method by introducing cascading dense connections in a very deep network. In detail, they construct the cascading dense network (CDN) to fully make use of the hierarchical features from all the convolutional layers, which implements a cascading mechanism upon the dense connected convolutional layers. The cascading dense connection in the CDN enables short and long paths to be built directly from the output to each layer, alleviating the vanishing‐gradient problem of very deep networks. Extensive experiments show that CDN achieves state‐of‐the‐art performance on traditional metrics (PSNR and SSIM). Also, they introduce object recognition as the additional evaluation metric for SISR, which further demonstrates the effectiveness of the authors’ method.

Recently, deep convolutional neural networks (CNNs) have been demonstrated excellent performance on single image super-resolution (SISR). However, most deep learning based methods lack the ability to distinguish features in network. For image super-resolution, it is important to design an effective prior to learn the correlation of various feature. To solve this problem, we propose a non-local hierarchical residual network (NHRN) for SISR. Specifically, we introduce a non-local module to measure the self-similarity between each pixels in the feature map, and obtain a weight matrix guiding the deep network to find more precise relationship between L-R and HR images. Thus our method reconstruct images with a sharper edge. In addition, we employ group convolutions to build a hierarchical residual structure, which enable the network extract image features hierarchically. It can reduce the executive time while ensuring reconstruct image quality. Extensive experiments show that our NHRN achieves better accuracy and speed against state-of-the-art methods.

MSAR-Net: Multi-scale attention based light-weight image super-resolution

Deep convolutional neural networks (CNNs) have recently achieved great success for single image super-resolution (SISR) task due to their powerful feature representation capabilities. The most recent deep learning based SISR methods focus on designing deeper / wider models to learn the non-linear mapping between low-resolution (LR) inputs and high-resolution (HR) outputs. These existing SR methods do not take into account the image observation (physical) model and thus require a large number of network's trainable parameters with a great volume of training data. To address these issues, we propose a deep Iterative Super-Resolution Residual Convolutional Network (ISRResCNet) that exploits the powerful image regularization and large-scale optimization techniques by training the deep network in an iterative manner with a residual learning approach. Extensive experimental results on various super-resolution benchmarks demonstrate that our method with a few trainable parameters improves the results for different scaling factors in comparison with the state-of-art methods.

Convolutional neural networks (CNNs) have recently made great progress in single image super-resolution (SISR) due to their powerful feature representation. However, most existing CNN-based SR methods mainly focus on designing deeper architectures to obtain superior representations, neglecting the global feature correlations inside networks, thus limiting the discriminative learning ability. In this paper, we propose a spatial graph convolutional network (SGCN) to amplify feature representation for high quality image rendering. Specifically, we present a residual feature refinement module (RFRM) in SGCN to encode correlations in both channel and spatial dimensions of features. A novel spatial graph attention (SGA) is deployed to model the correlations with awareness of global information for the feature enhancement. Further-more, we utilize the Gram matrix to infer the global relations in features, which adaptively measures the pixel-wise spatial attention with free parameters. Experimental results demonstrate the superiority of our SGCN over state-of-the-art SISR methods in terms of quantitative metrics and visual quality.

Recent research on single image super-resolution (SISR) using convolutional neural networks (CNNs) with the utilization of residual structures and attention mechanisms to utilize image features has demonstrated excellent performance. However, previous SISR techniques mainly integrated extracted image features within a deep or wide network architecture, ignoring the interaction between multiscale features and the diversity of features. At the same time, SISR is also a typical ill-posed problem in that it allows for several predictions for a given LR image. These problems limit the great learning ability of CNNs. To solve these problems, we propose a closed-loop residual attention network (CLRAN) to extract and interact with all the available diversity of features features efficiently and limit the space of possible function solutions. Specifically, we design an enhanced residual attention block (ERA) to extract features, and it dynamically assigns weight to the internal attention branches. The ERA combines multi-scale block (MSB) and enhanced attention mechanism (EAM) base on the residual module. The MSB adaptively detects multiscale image features of different scales by using different 3 × 3 convolution kernels. The EAM combines multi-spectral channel attention (MSCA) and spatial attention (SA). Therefore, the EAM extracts different frequency component information and spatial information to utilize the diversity features. Furthermore, we apply the progressive network architecture and learn an additional map for model monitoring, which forms a closed-loop with the mapping already learned by the LR to HR function. Extensive experiments demonstrate that our CLRAN outperforms the state-of-the-art SISR methods on public datasets for both ×4 and ×8, proving its accuracy and visual perception.