Super-Resolution by Image Enhancement Using Texture Transfer

We present a self-learning approach for single image super-resolution (SR), with the ability to preserve high frequency components such as edges in resulting high resolution (HR) images. Given a low-resolution (LR) input image, we construct its image pyramid and produce a super pixel dataset. By extracting context information from the super-pixels, we propose to deploy context-specific contour let transform on them in order to model the relationship (via support vector regression) between the input patches and their associated directional high-frequency responses. These learned models are applied to predict the SR output with satisfactory quality. Unlike prior learning-based SR methods, our approach advances a self-learning technique and does not require the self similarity of image patches within or across image scales. More importantly, we do not need to collect training LR/HR image data in advance and only require a single LR input image. Empirical results verify the effectiveness of our approach, which quantitatively and qualitatively outperforms existing interpolation or learning-based SR methods.

This paper presents a novel single-image super-resolution (SR) procedure, which upscales a given low-resolution (LR) input image to a high-resolution image while preserving the textural and structural information. First, we construct a new type of bivariate rational fractal interpolation model and investigate its analytical properties. This model has different forms of expression with various values of the scaling factors and shape parameters; thus, it can be employed to better describe image features than current interpolation schemes. Furthermore, this model combines the advantages of rational interpolation and fractal interpolation, and its effectiveness is validated through theoretical analysis. Second, we develop a single-image SR algorithm based on the proposed model. The LR input image is divided into texture and non-texture regions, and then, the image is interpolated according to the characteristics of the local structure. Specifically, in the texture region, the scaling factor calculation is the critical step. We present a method to accurately calculate scaling factors based on local fractal analysis. Extensive experiments and comparisons with the other state-of-the-art methods show that our algorithm achieves competitive performance, with finer details and sharper edges.

Super-resolution (SR) has become more vital, because of its capability to generate high-quality ultra-high definition (UHD) high-resolution (HR) images from low-resolution (LR) input images. Conventional SR methods entail high computational complexity, which makes them difficult to be implemented for up-scaling of full-high-definition input images into UHD-resolution images. Nevertheless, our previous super-interpolation (SI) method showed a good compromise between Peak-Signal-to-Noise Ratio (PSNR) performances and computational complexity. However, since SI only utilizes simple linear mappings, it may fail to precisely reconstruct HR patches with complex texture. In this paper, we present a novel SR method, which inherits the large-to-small patch conversion scheme from SI but uses global regression based on local linear mappings (GLM). Thus, our new SR method is called GLM-SI. In GLM-SI, each LR input patch is divided into 25 overlapped subpatches. Next, based on the local properties of these subpatches, 25 different local linear mappings are applied to the current LR input patch to generate 25 HR patch candidates, which are then regressed into one final HR patch using a global regressor. The local linear mappings are learned cluster-wise in our off-line training phase. The main contribution of this paper is as follows: Previously, linear-mapping-based conventional SR methods, including SI only used one simple yet coarse linear mapping to each patch to reconstruct its HR version. On the contrary, for each LR input patch, our GLM-SI is the first to apply a combination of multiple local linear mappings, where each local linear mapping is found according to local properties of the current LR patch. Therefore, it can better approximate nonlinear LR-to-HR mappings for HR patches with complex texture. Experiment results show that the proposed GLM-SI method outperforms most of the state-of-the-art methods, and shows comparable PSNR performance with much lower computational complexity when compared with a super-resolution method based on convolutional neural nets (SRCNN15). Compared with the previous SI method that is limited with a scale factor of 2, GLM-SI shows superior performance with average 0.79 dB higher in PSNR, and can be used for scale factors of 3 or higher.

Single image super resolution based on sparse domain selection

Resolution enhancement (RE) schemes (which are not based on wavelets) suffer from the drawback of losing high-frequency contents (which results in blurring). The discrete-wavelet-transform-based (DWT) RE scheme generates artifacts (due to a DWT shift-variant property). A wavelet-domain approach based on dual-tree complex wavelet transform (DT-CWT) and nonlocal means (NLM) is proposed for RE of the satellite images. A satellite input image is decomposed by DT-CWT (which is nearly shift invariant) to obtain high-frequency subbands. The high-frequency subbands and the low-resolution (LR) input image are interpolated using the Lanczos interpolator. The high-frequency subbands are passed through an NLM filter to cater for the artifacts generated by DT-CWT (despite of its nearly shift invariance). The filtered high-frequency subbands and the LR input image are combined using inverse DT-CWT to obtain a resolution-enhanced image. Objective and subjective analyses reveal superiority of the proposed technique over the conventional and state-of-the-art RE techniques.

Example learning-based image super-resolution (SR) is recognized as an effective way to produce a high-resolution (HR) image with the help of an external training set. The effectiveness of learning-based SR methods, however, depends highly upon the consistency between the supporting training set and low-resolution (LR) images to be handled. To reduce the adverse effect brought by incompatible high-frequency details in the training set, we propose a single image SR approach by learning multiscale self-similarities from an LR image itself. The proposed SR approach is based upon an observation that small patches in natural images tend to redundantly repeat themselves many times both within the same scale and across different scales. To synthesize the missing details, we establish the HR-LR patch pairs using the initial LR input and its down-sampled version to capture the similarities across different scales and utilize the neighbor embedding algorithm to estimate the relationship between the LR and HR image pairs. To fully exploit the similarities across various scales inside the input LR image, we accumulate the previous resultant images as training examples for the subsequent reconstruction processes and adopt a gradual magnification scheme to upscale the LR input to the desired size step by step. In addition, to preserve sharper edges and suppress aliasing artifacts, we further apply the nonlocal means method to learn the similarity within the same scale and formulate a nonlocal prior regularization term to well pose SR estimation under a reconstruction-based SR framework. Experimental results demonstrate that the proposed method can produce compelling SR recovery both quantitatively and perceptually in comparison with other state-of-the-art baselines.

Positron Emission Tomography (PET) is a nuclear, in-vivo medical imaging technology which can make 3D images of tissue metabolic act, in which high dose of tracer is needed to obtain a high quality PET image, which affects patients' health. Due to the limitations of using high dose tracer and limitations of physical imaging systems, it is not easy to get an image in desired resolution. Simplest approach to generate a high resolution image is by post processing. Single Image Super Resolution (SISR) is a post processing procedure to retrieve a high resolution image from a low resolution input. We propose a convolutional neural network trained on PET images which can estimate a high resolution PET image from its input low resolution image.

The improvement of the spatial resolution of natural images is an important task for many practical applications of current image processing. The endeavor of applying single image super-resolution, which only uses one low-resolution input image, is frequently hampered by the presence of noise or artifacts in the input image, so denoising and super-resolution algorithms are usually applied to obtain a noiseless high-resolution image. In this work, several traditional and deep learning methods for denoising and super-resolution are hybridized to ascertain which combinations of them yield the best possible quality. The experimental design includes the introduction of Gaussian noise, Poisson noise, salt-and-pepper noise, and uniform noise into the low-resolution inputs. It is found that denoising must be carried out before super-resolution for the best results. Moreover, the obtained results indicate that deep learning techniques clearly outperform the traditional ones.

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

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.

Most of the previous models performed well for Single Image Super-Resolution (SISR). In these methods, the Low Resolution (LR) input image is amplified to the size of High Resolution (HR) through bicubic interpolation. However, bicubic interpolation can not represent the high frequency features of images with only one filter. Therefore, in this paper, we used a original framework which can effectively extract the feature maps from the input image space and transform to HR feature maps based on Spatial Transformer Networks (STN). In our STN-SR method, there are three kinds of parameters should be learned from the model: (i) a serial of filters to extract LR image feature maps; (ii)a local small network to learn parameters of the transformation $\Gamma_{\theta}(G)$ and (iii) the filter parameters to restore the HR patchs from the input HR feature maps through a restoring layer. Our model directly focus on the whole image, the proposed STN-SR method does not clip the image into many small size patches, and can use the image gobal message to rebuild more robust local texture. Compared to privious SR methods, the proposed STN-SR method can gain completely real image, while illustrating better edge and texture preservation performance.

Learning based single image super-resolution (SISR) for real-world images has been an active research topic yet a challenging task, due to the lack of paired low-resolution (LR) and high-resolution (HR) training images. Most of the existing unsupervised real-world SISR methods adopt a two-stage training strategy by synthesizing realistic LR images from their HR counterparts first, then training the super-resolution (SR) models in a supervised manner. However, the training of image degradation and SR models in this strategy are separate, ignoring the inherent mutual dependency between downscaling and its inverse upscaling process. Additionally, the ill-posed nature of image degradation is not fully considered. In this paper, we propose an image downscaling and SR model dubbed as SDFlow, which simultaneously learns a bidirectional many-to-many mapping between real-world LR and HR images unsupervisedly. The main idea of SDFlow is to decouple image content and degradation information in the latent space, where content information distribution of LR and HR images is matched in a common latent space. Degradation information of the LR images and the high-frequency information of the HR images are fitted to an easy-to-sample conditional distribution. Experimental results on real-world image SR datasets indicate that SDFlow can generate diverse realistic LR and SR images both quantitatively and qualitatively.

In this study, a novel single image super-resolution (SR) method, which uses a generated dictionary from pairs of high-resolution (HR) images and their corresponding low-resolution (LR) representations, is proposed. First, HR and LR dictionaries are created by dividing HR and LR images into patches Afterwards, when performing SR, the distance between every patch of the input LR image and those of available LR patches in the LR dictionary are calculated. The minimum distance between the input LR patch and those in the LR dictionary is taken, and its counterpart from the HR dictionary will be passed through an illumination enhancement process resulting in consistency of illumination between neighbour patches. This process is applied to all patches of the LR image. Finally, in order to remove the blocking effect caused by merging the patches, an average of the obtained HR image and the interpolated image is calculated. Furthermore, it is shown that the stabe of dictionaries is reducible to a great degree. The speed of the system is improved by 62.5%. The quantitative and qualitative analyses of the experimental results show the superiority of the proposed technique over the conventional and state-of-the-art methods.

SRDiff: Single image super-resolution with diffusion probabilistic models

Image super-resolution (SR) produces a high-resolution (HR) image using either a single or multiple low-resolution (LR) input image(s) of the same scene. Sparse representation techniques are effectively applied for image SR because of their high reconstruction accuracy. SR reconstruction is accomplished through learning of an overcomplete dictionary and then solving a series of regularization problems applied on each patch extracted from the input image. This paper demonstrates a sparse representation based coupled overcomplete dictionary training and SR procedure for LR multispectral images. The proposed work is also implemented using a multicore parallel processing technique to provide faster reconstruction. Experimental results show superiority of the proposed method over some others.