Single image super-resolution reconstruction via combination mapping with sparse coding

Sparse representation is one of the effective approaches to perform the feature extraction on the high-dimensional signals and has been widely used in the field of image processing. In this paper, a novel single image super-resolution (SR) reconstruction algorithm based on sparse coding is presented. Specifically, due to the fact that sparse coding has the potential to greatly facilitate the biological-neural system and to deal with a large amount of complex data, while consuming a finite amount of energy, a memristor crossbar array synthesized by a threshold-type memristor model is constructed for the implementation of a hardware-friendly sparse coding method, i.e., the Locally Competitive Algorithm (LCA). Meanwhile, the relevant online dictionary learning method is provided, along with the description of weight programming strategy. Furthermore, the specific description of the image SR via the LCA sparse coding is provided. For the verification purpose, a series of experiments are carried out to illustrate the validity and effectiveness of the entire scheme.

Deep learning can be used to achieve single‐image superresolution (SR) reconstruction. To address problems encountered during this process, such as the number of network parameters, high training requirements on equipment performance, and inability to downsample certain SR images accurately, an image SR reconstruction algorithm based on deep residual network optimization is proposed. The model introduces wavelet transforms based on the original U‐Net, where the U‐Net is trained to obtain SR wavelet feature images at multiple scales simultaneously. This approach reduces the mapping space for the network to learn low‐ to high‐resolution image mapping, which in turn reduces the training difficulty of the model. In terms of network details, the inverse wavelet transform is used in image upsampling to enhance the sparsity of the reconstruction layer in the original network. The network structure of the U‐Net upsampling is adjusted slightly to enable the network to distinguish wavelet images from feature images, thereby improving the richness of the features extracted by the model. The experimental results show that the peak signal‐to‐noise ratio (PSNR) of the fourfold SR model is 32.35 and 28.68 on the Set5 and Set14 validation sets, respectively. Compared with networks that use wavelet prediction mechanisms, such as the deep wavelet prediction SR (DWSR) and deep wavelet prediction‐based residual SR (DWRSR) models, the PSNR for all the tested public datasets is improved by 0.5. The method yields superior results in terms of both visual effect and PSNR, demonstrating the feasibility of the wavelet prediction mechanism in SR reconstruction and thus offering application value and research significance.

Methods based on sparse coding have been successfully used in single-image superresolution (SR) reconstruction. However, the traditional sparse representation-based SR image reconstruction for infrared (IR) images usually suffers from three problems. First, IR images always lack detailed information. Second, a traditional sparse dictionary is learned from patches with a fixed size, which may not capture the exact information of the images and may ignore the fact that images naturally come at different scales in many cases. Finally, traditional sparse dictionary learning methods aim at learning a universal and overcomplete dictionary. However, many different local structural patterns exist. One dictionary is inadequate in capturing all of the different structures. We propose a novel IR image SR method to overcome these problems. First, we combine the information from multisensors to improve the resolution of the IR image. Then, we use multiscale patches to represent the image in a more efficient manner. Finally, we partition the natural images into documents and group such documents to determine the inherent topics and to learn the sparse dictionary of each topic. Extensive experiments validate that using the proposed method yields better results in terms of quantitation and visual perception than many state-of-the-art algorithms.

Convolutional neural networks have recently demonstrated high-quality reconstruction for single image super-resolution. However, existing methods often require a large number of network parameters and entail heavy computational loads at runtime for generating high-accuracy super-resolution results. In this paper, we propose the deep Laplacian Pyramid Super-Resolution Network for fast and accurate image super-resolution. The proposed network progressively reconstructs the sub-band residuals of high-resolution images at multiple pyramid levels. In contrast to existing methods that involve the bicubic interpolation for pre-processing (which results in large feature maps), the proposed method directly extracts features from the low-resolution input space and thereby entails low computational loads. We train the proposed network with deep supervision using the robust Charbonnier loss functions and achieve high-quality image reconstruction. Furthermore, we utilize the recursive layers to share parameters across as well as within pyramid levels, and thus drastically reduce the number of parameters. Extensive quantitative and qualitative evaluations on benchmark datasets show that the proposed algorithm performs favorably against the state-of-the-art methods in terms of run-time and image quality.

Recently, Convolutional Neural Network (CNN) ba-sed methods have demonstrated high-quality reconstruction for Single Image Super-Resolution (SISR). Particularly, resi-dual learning shows improved performance. For the SR recon-struction performance of the CNN-based models, the design of model architecture is very important. In this paper, we present SRDPN, a super-resolution network based on Dual Path Net-work (DPN). The DPN combines the advantages of the state-of-the-art Residual Network (ResNet) and Dense Convolu-tional Network (DenseNet), which enjoys lower computational cost, lower memory consumption and higher parameter effici-ency. Extensive qualitative and quantitative evaluations on benchmark datasets show that our proposed method performs better than the state-of-the-art methods in terms of peak signal-to-noise ratio (PSNR) and image quality.

Single image super-resolution (SISR) reconstruction for magnetic resonance imaging (MRI) has generated significant interest because of its potential to not only speed up imaging but to improve quantitative processing and analysis of available image data. Generative Adversarial Networks (GAN) have proven to perform well in image recovery tasks. In this work, we followed the GAN framework and developed a generator coupled with discriminator to tackle the task of 3D SISR on T1 brain MRI images. We developed a novel 3D memory-efficient residual-dense block generator (MRDG) that achieves state-of-the-art performance in terms of SSIM (Structural Similarity), PSNR (Peak Signal to Noise Ratio) and NRMSE (Normalized Root Mean Squared Error) metrics. We also designed a pyramid pooling discriminator (PPD) to recover details on different size scales simultaneously. Finally, we introduced model blending, a simple and computational efficient method to balance between image and texture quality in the final output, to the task of SISR on 3D images.

Single image super resolution based on sparse domain selection

High-resolution (HR) remote sensing imagery plays a critical role in remote sensing image interpretation, and single image super-resolution (SISR) reconstruction technology is becoming increasingly valuable and significant. The state-of-the-art deep-learning-based SISR methods have demonstrated remarkable advantages, while reconstructing complex texture details still remains a big challenge. Besides, as a typical ill-posed inverse problem, how to determine the optimal solution is another important topic. To address these problems, in this work, a dual learning-based graph neural network (DLGNN) is proposed, in which the GNN is utilized to consider the self-similarity patches in remote sensing imagery by aggregating cross-scale neighboring feature patches, and dual learning strategy is adopted to refine the reconstruction results by constraining the mapping process in terms of the loss function, transferring the typical ill-posed problem to a well-posed one. Abundant experiments on 3K VEHICLE_SR datasets and Massachusetts Roads demonstrate the validity and outstanding performance for remote sensing image super-resolution tasks compared with other state-of-the-art super-resolution construction methods. Code is available at https://github.com/CUG-RS/DLGNN.

The single image super resolution recovers missing high resolution details so as to reconstruct a high resolution image from a single low resolution image. This paper proposes a novel directionally adaptive, learning-based, single image super resolution method using multiple direction wavelet transform, called directionlets. Here, critically sampled directionlets are used to capture directional features effectively and to extract edge information along different directions from a set of available high resolution images. This information is used as the training set for super resolving a low resolution input image. The directionlet coefficients at finer scales of its high resolution image are learned locally from this training set and the inverse directionlet transform recovers the super resolved high resolution image. The simulation results showed that the proposed directionlet approach outperforms standard interpolation techniques like cubic spline interpolation as well as standard wavelet-based learning, both visually and in terms of the mean squared error (MSE) values. The SNR scores for cubic spline interpolation, wavelet and directionlet method are 13.6998 dB, 23.8324 dB and 30.8654 dB respectively for Barbara.

Recent studies have shown that the use of deep convolutional neural networks (Deep CNNs) can significantly improve the performance of single-image super-resolution reconstruction. In this paper, we propose a highly accurate and fast single-image super-resolution reconstruction (SISR) method by introducing dense skip connections and Inception-ResNet in deep convolutional neural networks. In the proposed network, all previous layer feature maps are used as input for each subsequent layer to promote feature reuse and alleviate the vanishing-gradient problem. In addition, the parallelized CNNs structure used reduces the size of the output of the previous layer, thereby accelerating the calculation speed and reducing the feature loss. Moreover, we only learn the residuals between the high-resolution images and the low-resolution images, and use the adjustable gradient cropping to achieve a very high learning rate. Results on benchmark datasets demonstrate that the proposed model not only achieves higher accuracy but also enables faster and more efficient calculations than the state-of-the-art methods.

The resolution of remote-sensing images is directly related to the application value of data products. Due to differences in imaging characteristics between digital cameras and remote-sensing cameras, the existing network models cannot get the best super resolution (SR) results of satellite images. In response to the requirements of remote-sensing image production, we propose a single image super-resolution (SISR) reconstruction method for specific type of remote-sensing satellite. First, we measure and model the imaging degradation phenomenon of remote-sensing satellite, including the image blur and noise model. Then, high-quality aerial images are down-sampled and degraded to construct paired training image datasets. We chose the most popular Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN) as the basic structure and optimized the number of Residual-in-Residual Dense Block (RRDB) modules to further improve the processing efficiency. Finally, we perform a series of quantitative measurements of the SR image results, including image interpretation capability, reconstruction accuracy, ground resolution distance, and data processing efficiency, using higher resolution remote-sensing images as the benchmark. The experimental results demonstrate that our proposed method has higher interpretation capability and reconstruction accuracy for the SR processing of specific type remote-sensing satellite. Our proposed method is evaluated within a real satellite image product, that demonstrated it has the capability of pipeline super-resolution processing of remote sensing data products.

This chapter concentrates the problem of recovery a high-resolution (HR) image from a single low-resolution input image. Recent research proposed to deal with the image super-resolution problem with sparse coding, which is based on the well reconstruction of any local image patch by a sparse linear combination of an appropriately chosen over-complete dictionary. Therein the chosen LR (Low-resolution) and HR (High-resolution) dictionaries have to be exactly corresponding for well reconstructing the local image patterns. However, the conventional sparse coding based image super-resolution usually achieves a global dictionary D=[D l ; D h ] by jointly training the concatenated LR and HR local image patches, and then reconstruct the LR and HR image as a linear combination of the separated D l and D h . This strategy only can achieve the global minimum reconstructing error of LR and HR local patches, and usually cannot obtain the exactly corresponding LR and HR dictionaries. In addition, the accurate coefficients for reconstructing the HR image patch using HR dictionary D h are also unable to be estimated using only the LR input and the LR dictionary D l . Therefore, this paper proposes to firstly learn the HR dictionary D h from the features of the training HR local patches, and then propagates the HR dictionary to the LR one, called as HR2LR dictionary propagation, by mathematical proving and statistical analysis. The effectiveness of the proposed HR2LR dictionary propagation in sparse coding for super-resolution is demonstrated by comparison with the conventional super-resolution approaches such as sparse coding and interpolation.

Recent studies have shown that the use of deep convolutional neural networks (CNNs) can improve the performance of single image super-resolution reconstruction (SISR) methods. However, the existing CNN-based SISR model ignores the multi-scale features and shallow and deep features of the image, resulting in relatively low image reconstruction performance. To address these issues, this paper proposes a new multi-scale fractal residual network (MSFRN) for image super-resolution. On the basis of residual learning, a multi-scale fractal residual block (MSFRB) is designed. This block uses convolution kernels of different sizes to extract image multi-scale features and uses multiple paths to extract and fuse image features of different depths. Then, the shallow features extracted at the shallow feature extraction stage and the local features output by all MSFRBs are used to perform global hierarchical feature fusion. Finally, through sub-pixel convolution, the fused global features are used to reconstruct high-resolution images from low-resolution images. The experimental results on the five standard benchmark datasets show that MSFRN improved subjective visual effects and objective image quality evaluation indicators, and is superior to other state-of-the-art SISR methods.

High-quality reference images can be seen as a baseline for many image reconstruction techniques. The reconstruction performance gets improved as shown in the Reference Image Constrained Regularization by Denoising. But, low-quality reference images lead to disappointing results, even underperforming methods that use no reference at all. To address this limitation, a novel framework termed Adaptive Reference Constrained Regularization by Denoising is proposed in this paper. This new method mitigates the influence of low-quality reference images by adaptively leveraging the reconstruction results obtained from the existing algorithm to form a more reliable constraint. Furthermore, an enhanced version is introduced, equipped with a self-tuning parameter that progressively diminishes the weight of a poor reference throughout the optimization iterations. Extensive experiments on single-image super-resolution and MRI reconstruction demonstrate that the proposed approach significantly outperforms state-of-the-art methods, including Regularization by Denoising and Reference Image Constrained Regularization by Denoising, in both quantitative metrics and visual fidelity.

In order to improve the super-resolution reconstruction effect of the single image, a novel multiple dictionaries learning via support vector regression (SVR) and improved iterative back-projection (IBP) are proposed. To characterize the image structure, the low-frequency dictionary is constructed from the normalized brightness of low-frequency image patches in a discrete-cosine-transform (DCT) domain. Pixels determined by Gaussian weighting are added to the input vector to restore more high-frequency information when training the high-frequency image patch dictionary in the space domain. During post-processing, the improved IBP is employed to reduce regression errors each time. Experiment results show that the peak signal-to-noise ratio (PSNR)and structural similarity (SSIM) of the proposed method are enhanced by 1.6%–5.5% and 1.5%–13.1% compared with those of bicubic interpolation, and the proposed method visually outperforms several algorithms.