Learning multiple linear mappings for efficient single image super-resolution

Sparse data-based image super-resolution with ANFIS interpolation

Image processing is a very broad field containing various areas, including image super-resolution (ISR) which re-represents a low-resolution image as a high-resolution one through a certain means of image transformation. The problem with most of the existing ISR methods is that they are devised for the condition in which sufficient training data is expected to be available. This article proposes a new approach for sparse data-based (rather than sufficient training data-based) ISR, by the use of an ANFIS (Adaptive Network-based Fuzzy Inference System) interpolation technique. Particularly, a set of given image training data is split into various subsets of sufficient and sparse training data subsets. Typical ANFIS training process is applied for those subsets involving sufficient data, and ANFIS interpolation is employed for the rest that contains sparse data only. Inadequate work is available in the current literature for the sparse data-based ISR. Consequently, the implementations of the proposed sparse data-based approach, for both training and testing processes, are compared with the state-of-the-art sufficient data-based ISR methods. This is of course very challenging, but the results of experimental evaluation demonstrate positively about the efficacy of the work presented herein.

Image Denoising Using a Novel Deep Generative Network with Multiple Target Images and Adaptive Termination Condition

Image denoising, a classic ill-posed problem, aims to recover a latent image from a noisy measurement. Over the past few decades, a considerable number of denoising methods have been studied extensively. Among these methods, supervised deep convolutional networks have garnered increasing attention, and their superior performance is attributed to their capability to learn realistic image priors from a large amount of paired noisy and clean images. However, if the image to be denoised is significantly different from the training images, it could lead to inferior results, and the networks may even produce hallucinations by using inappropriate image priors to handle an unseen noisy image. Recently, deep image prior (DIP) was proposed, and it overcame this drawback to some extent. The structure of the DIP generator network is capable of capturing the low-level statistics of a natural image using an unsupervised method with no training images other than the image itself. Compared with a supervised denoising model, the unsupervised DIP is more flexible when processing image content that must be denoised. Nevertheless, the denoising performance of DIP is usually inferior to the current supervised learning-based methods using deep convolutional networks, and it is susceptible to the over-fitting problem. To solve these problems, we propose a novel deep generative network with multiple target images and an adaptive termination condition. Specifically, we utilized mainstream denoising methods to generate two clear target images to be used with the original noisy image, enabling better guidance during the convergence process and improving the convergence speed. Moreover, we adopted the noise level estimation (NLE) technique to set a more reasonable adaptive termination condition, which can effectively solve the problem of over-fitting. Extensive experiments demonstrated that, according to the denoising results, the proposed approach significantly outperforms the original DIP method in tests on different databases. Specifically, the average peak signal-to-noise ratio (PSNR) performance of our proposed method on four databases at different noise levels is increased by 1.90 to 4.86 dB compared to the original DIP method. Moreover, our method achieves superior performance against state-of-the-art methods in terms of popular metrics, which include the structural similarity index (SSIM) and feature similarity index measurement (FSIM). Thus, the proposed method lays a good foundation for subsequent image processing tasks, such as target detection and super-resolution.

PSR: Unified Framework of Parameter-Learning-Based MR Image Superresolution

Magnetic resonance imaging has significant applications for disease diagnosis. Due to the particularity of its imaging mechanism, hardware imaging suffers from resolution and reaches its limit, and higher radiation intensity and longer radiation time will cause damage to the human body. The problem is expected to be solved by a superresolution algorithm, especially the image superresolution based on sparse reconstruction has good performance. Dictionary generation is a key issue that affects the performance of superresolution algorithms, and dictionary performance is affected by dictionary construction parameters: balance parameters, dictionary size, overlapping block size, and a number of training sample blocks. In response to this problem, we propose an optimal dictionary construction parameter search method through the experiment to find the optimal dictionary construction parameters on the MR image and compare them with the dictionary obtained by multiple sets of random dictionary construction parameters. The dictionary we searched for the optimal parameters of the dictionary construction training has more powerful feature expressions, which can improve the superresolution effect of MR images.

Resolution enhancement and realistic speckle recovery with generative adversarial modeling of micro-optical coherence tomography

A resolution enhancement technique for optical coherence tomography (OCT), based on Generative Adversarial Networks (GANs), was developed and investigated. GANs have been previously used for resolution enhancement of photography and optical microscopy images. We have adapted and improved this technique for OCT image generation. Conditional GANs (cGANs) were trained on a novel set of ultrahigh resolution spectral domain OCT volumes, termed micro-OCT, as the high-resolution ground truth (∼1 μm isotropic resolution). The ground truth was paired with a low-resolution image obtained by synthetically degrading resolution 4x in one of (1-D) or both axial and lateral axes (2-D). Cross-sectional image (B-scan) volumes obtained from in vivo imaging of human labial (lip) tissue and mouse skin were used in separate feasibility experiments. Accuracy of resolution enhancement compared to ground truth was quantified with human perceptual accuracy tests performed by an OCT expert. The GAN loss in the optimization objective, noise injection in both the generator and discriminator models, and multi-scale discrimination were found to be important for achieving realistic speckle appearance in the generated OCT images. The utility of high-resolution speckle recovery was illustrated by an example of micro-OCT imaging of blood vessels in lip tissue. Qualitative examples applying the models to image data from outside of the training data distribution, namely human retina and mouse bladder, were also demonstrated, suggesting potential for cross-domain transferability. This preliminary study suggests that deep learning generative models trained on OCT images from high-performance prototype systems may have potential in enhancing lower resolution data from mainstream/commercial systems, thereby bringing cutting-edge technology to the masses at low cost.

Single image super-resolution via Image Quality Assessment-Guided Deep Learning Network

In recent years, deep learning (DL) networks have been widely used in super-resolution (SR) and exhibit improved performance. In this paper, an image quality assessment (IQA)-guided single image super-resolution (SISR) method is proposed in DL architecture, in order to achieve a nice tradeoff between perceptual quality and distortion measure of the SR result. Unlike existing DL-based SR algorithms, an IQA net is introduced to extract perception features from SR results, calculate corresponding loss fused with original absolute pixel loss, and guide the adjustment of SR net parameters. To solve the problem of heterogeneous datasets used by IQA and SR networks, an interactive training model is established via cascaded network. We also propose a pairwise ranking hinge loss method to overcome the shortcomings of insufficient samples during training process. The performance comparison between our proposed method with recent SISR methods shows that the former achieves a better tradeoff between perceptual quality and distortion measure than the latter. Extensive benchmark experiments and analyses also prove that our method provides a promising and opening architecture for SISR, which is not confined to a specific network model.

Enhanced Non-Local Total Variation Model and Multi-Directional Feature Prediction Prior for Single Image Super Resolution

It is widely acknowledged that single image super-resolution (SISR) methods play a critical role in recovering the missing high-frequencies in an input low-resolution image. As SISR is severely ill-conditioned, image priors are necessary to regularize the solution spaces and generate the corresponding high-resolution image. In this paper, we propose an effective SISR framework based on the enhanced non-local similarity modeling and learning-based multi-directional feature prediction (ENLTV-MDFP). Since both the modeled and learned priors are exploited, the proposed ENLTV-MDFP method benefits from the complementary properties of the reconstruction-based and learning-based SISR approaches. Specifically, for the non-local similarity-based modeled prior [enhanced non-local total variation, (ENLTV)], it is characterized via the decaying kernel and stable group similarity reliability schemes. For the learned prior [multi-directional feature prediction prior, (MDFP)], it is learned via the deep convolutional neural network. The modeled prior performs well in enhancing edges and suppressing visual artifacts, while the learned prior is effective in hallucinating details from external images. Combining these two complementary priors in the MAP framework, a combined SR cost function is proposed. Finally, the combined SR problem is solved via the split Bregman iteration algorithm. Based on the extensive experiments, the proposed ENLTV-MDFP method outperforms many state-of-the-art algorithms visually and quantitatively.

Dictionary optimization and constraint neighbor embedding-based dictionary mapping for superdimension reconstruction of porous media

The three-dimensional (3D) structure of a digital core can be reconstructed from a single two-dimensional (2D) image via mathematical modeling. In classical mathematical modeling algorithms, such as multipoint geostatistics algorithms, the optimization of pattern sets (dictionaries) and the mapping problems are important issues. However, they have rarely been discussed thus far. Pattern set (dictionary)-related problems include the pattern set (dictionary) size problem and the one-to-many mapping problem in a pattern set (dictionary). The former directly affects the completeness of the dictionary, while the latter is manifested such that a single to-be-matched 2D patch has multiple matching patterns in the library and it is hence necessary to select these modes to establish an optimal mapping relationship. Whether the two above-mentioned problems can be properly resolved is directly related to the accuracy of the reconstruction results. Super-dimension reconstruction is a new 3D reconstruction method proposed by introducing the concepts of training dictionary, prior model, and mapping into the reconstruction of the digital core from the field of super-resolution reconstruction. In addition, mapping relationship extraction and dictionary building are also key issues in super-dimension reconstruction. Therefore, this paper discusses these common dictionary-related problems from the perspective of super-dimension dictionaries. We propose dictionary optimization using augmentation dictionaries and clustering based on the boundary features of the dictionary elements to improve the completeness and expand the expression ability of the dictionary. Furthermore, we propose constraint neighbor embedding-based dictionary mapping to establish a more reasonable dictionary mapping relationship for super-dimension reconstruction, and we solve the one-to-many mapping problem in the dictionary. Our experimental results show that the performance of the super-dimension dictionary can be improved by the above-mentioned algorithm. Thus, through the optimized dictionary structure and mapping relationship determined by the above-mentioned methods, the 2D patch to be reconstructed can match a more accurate 3D block in the dictionary. Consequently, the reconstruction precision is improved.

Super-resolution imaging and field of view extension using a single camera with Risley prisms

A novel imaging method using Risley prisms is proposed to achieve super-resolution imaging and field of view (FOV) extension. The mathematical models are developed, and the solutions to sub-pixel imaging for super-resolution reconstruction are presented. Simulations show that the proposed method can enhance the image resolution up to optical diffraction limit of the optical system for imaging systems whose resolution is limited by pixel size. A prototype is developed. Experimental results show that the scene resolving capacity can be enhanced by 2.0 times with a resolution improvement factor of 4, and the FOV extension results accord with the simulations, providing a promising approach for super-resolution reconstruction, large FOV imaging, and foveated imaging with low cost and high efficiency.

Single Image Super-Resolution via Multiple Mixture Prior Models

Example learning-based single image super-resolution (SR) is a promising method for reconstructing a high-resolution (HR) image from a single-input low-resolution (LR) image. Lots of popular SR approaches are more likely either time-or space-intensive, which limit their practical applications. Hence, some research has focused on a subspace view and delivered state-of-the-art results. In this paper, we utilize an effective way with mixture prior models to transform the large nonlinear feature space of LR images into a group of linear subspaces in the training phase. In particular, we first partition image patches into several groups by a novel selective patch processing method based on difference curvature of LR patches, and then learning the mixture prior models in each group. Moreover, different prior distributions have various effectiveness in SR, and in this case, we find that student-t prior shows stronger performance than the well-known Gaussian prior. In the testing phase, we adopt the learned multiple mixture prior models to map the input LR features into the appropriate subspace, and finally reconstruct the corresponding HR image in a novel mixed matching way. Experimental results indicate that the proposed approach is both quantitatively and qualitatively superior to some state-of-the-art SR methods.

Hallucinating Face Image by Regularization Models in High-Resolution Feature Space

In this paper, we propose two novel regularization models in patch-wise and pixel-wise respectively, which are efficient to reconstruct high-resolution (HR) face image from low-resolution (LR) input. Unlike the conventional patch-based models which depend on the assumption of local geometry consistency in LR and HR spaces, the proposed method directly regularizes the relationship between the target patch and corresponding training set in the HR space. It avoids to deal with the tough problem of preserving local geometry in various resolutions. Taking advantage of kernel function in efficiently describing intrinsic features, we further conduct the patch-based reconstruction model in the high-dimensional kernel space for capturing nonlinear characteristics. Meanwhile, a pixel-based model is proposed to regularize the relationship of pixels in the local neighborhood, which can be employed to enhance the fuzzy details in the target HR face image. It privileges the reconstruction of pixels along the dominant orientation of structure, which is useful for preserving high-frequency information on complex edges. Finally, we combine the two reconstruction models into a unified framework. The output HR face image can be finally optimized by performing an iterative procedure. Experimental results demonstrate that the proposed face hallucination method produces superior performance than the state-of-the-art methods.

How Does the Low-Rank Matrix Decomposition Help Internal and External Learnings for Super-Resolution

Wisely utilizing the internal and external learning methods is a new challenge in super-resolution problem. To address this issue, we analyze the attributes of two methodologies and find two observations of their recovered details: 1) they are complementary in both feature space and image plane and 2) they distribute sparsely in the spatial space. These inspire us to propose a low-rank solution which effectively integrates two learning methods and then achieves a superior result. To fit this solution, the internal learning method and the external learning method are tailored to produce multiple preliminary results. Our theoretical analysis and experiment prove that the proposed low-rank solution does not require massive inputs to guarantee the performance, and thereby simplifying the design of two learning methods for the solution. Intensive experiments show the proposed solution improves the single learning method in both qualitative and quantitative assessments. Surprisingly, it shows more superior capability on noisy images and outperforms state-of-the-art methods.

Sparsity-Based Color Image Super Resolution via Exploiting Cross Channel Constraints

Sparsity constrained single image super-resolution (SR) has been of much recent interest. A typical approach involves sparsely representing patches in a low-resolution (LR) input image via a dictionary of example LR patches, and then using the coefficients of this representation to generate the high-resolution (HR) output via an analogous HR dictionary. However, most existing sparse representation methods for SR focus on the luminance channel information and do not capture interactions between color channels. In this paper, we extend sparsity-based SR to multiple color channels by taking the color information into account. Edge similarities amongst RGB color bands are exploited as cross channel correlation constraints. These additional constraints lead to a new optimization problem, which is not easily solvable; however, a tractable solution is proposed to solve it efficiently. Moreover, to fully exploit the complementary information among color channels, a dictionary learning method is also proposed specifically to learn color dictionaries that encourage edge similarities. Merits of the proposed method over state of the art are demonstrated both visually and quantitatively using image quality metrics.

Learning the Image Processing Pipeline

Many creative ideas are being proposed for image sensor designs, and these may be useful in applications ranging from consumer photography to computer vision. To understand and evaluate each new design, we must create a corresponding image processing pipeline that transforms the sensor data into a form, that is appropriate for the application. The need to design and optimize these pipelines is time-consuming and costly. We explain a method that combines machine learning and image systems simulation that automates the pipeline design. The approach is based on a new way of thinking of the image processing pipeline as a large collection of local linear filters. We illustrate how the method has been used to design pipelines for novel sensor architectures in consumer photography applications.

Dictionary learning based noisy image super-resolution via distance penalty weight model

In this study, we address the problem of noisy image super-resolution. Noisy low resolution (LR) image is always obtained in applications, while most of the existing algorithms assume that the LR image is noise-free. As to this situation, we present an algorithm for noisy image super-resolution which can achieve simultaneously image super-resolution and denoising. And in the training stage of our method, LR example images are noise-free. For different input LR images, even if the noise variance varies, the dictionary pair does not need to be retrained. For the input LR image patch, the corresponding high resolution (HR) image patch is reconstructed through weighted average of similar HR example patches. To reduce computational cost, we use the atoms of learned sparse dictionary as the examples instead of original example patches. We proposed a distance penalty model for calculating the weight, which can complete a second selection on similar atoms at the same time. Moreover, LR example patches removed mean pixel value are also used to learn dictionary rather than just their gradient features. Based on this, we can reconstruct initial estimated HR image and denoised LR image. Combined with iterative back projection, the two reconstructed images are applied to obtain final estimated HR image. We validate our algorithm on natural images and compared with the previously reported algorithms. Experimental results show that our proposed method performs better noise robustness.

Scalable Image Coding Based on Epitomes

In this paper, we propose a novel scheme for scalable image coding based on the concept of epitome. An epitome can be seen as a factorized representation of an image. Focusing on spatial scalability, the enhancement layer of the proposed scheme contains only the epitome of the input image. The pixels of the enhancement layer not contained in the epitome are then restored using two approaches inspired from local learning-based super-resolution methods. In the first method, a locally linear embedding model is learned on base layer patches and then applied to the corresponding epitome patches to reconstruct the enhancement layer. The second approach learns linear mappings between pairs of co-located base layer and epitome patches. Experiments have shown that the significant improvement of the rate-distortion performances can be achieved compared with the Scalable extension of HEVC (SHVC).

Single Image Super-Resolution Using Gaussian Process Regression With Dictionary-Based Sampling and Student- ${t}$ Likelihood

Gaussian process regression (GPR) is an effective statistical learning method for modeling non-linear mapping from an observed space to an expected latent space. When applying it to example learning-based super-resolution (SR), two outstanding issues remain. One is its high computational complexity restricts SR application when a large data set is available for learning task. The other is that the commonly used Gaussian likelihood in GPR is incompatible with the true observation model for SR reconstruction. To alleviate the above two issues, we propose a GPR-based SR method by using dictionary-based sampling (DbS) and student-t likelihood. Considering that dictionary atoms effectively span the original training sample space, we adopt a DbS strategy by combining all the neighborhood samples of each atom into a compact representative training subset so as to reduce the computational complexity. Based on statistical tests, we statistically validate that student-t likelihood is more suitable to build the observation model for the SR problem. Extensive experimental results show that the proposed method outperforms other competitors and produces more pleasing details in texture regions.

7T-guided super-resolution of 3T MRI

PURPOSE

High-resolution MR images can depict rich details of brain anatomical structures and show subtle changes in longitudinal data. 7T MRI scanners can acquire MR images with higher resolution and better tissue contrast than the routine 3T MRI scanners. However, 7T MRI scanners are currently more expensive and less available in clinical and research centers. To this end, we propose a method to generate super-resolution 3T MRI that resembles 7T MRI, which is called as 7T-like MR image in this paper.

METHODS

First, we propose a mapping from 3T MRI to 7T MRI space, using regression random forest. The mapped 3T MR images serve as intermediate results with similar appearance as 7T MR images. Second, we predict the final higher resolution 7T-like MR images based on sparse representation, using paired local dictionaries for both the mapped 3T MR images and 7T MR images.

RESULTS

Based on 15 subjects with both 3T and 7T MR images, the predicted 7T-like MR images by our method can best match the ground-truth 7T MR images, compared to other methods. Meanwhile, the experiment on brain tissue segmentation shows that our 7T-like MR images lead to the highest accuracy in the segmentation of WM, GM, and CSF brain tissues, compared to segmentations of 3T MR images as well as the reconstructed 7T-like MR images by other methods.

CONCLUSIONS

We propose a novel method for prediction of high-resolution 7T-like MR images from low-resolution 3T MR images. Our predicted 7T-like MR images demonstrate better spatial resolution compared to 3T MR images, as well as prediction results by other comparison methods. Such high-quality 7T-like MR images could better facilitate disease diagnosis and intervention.

Fast single image super-resolution using estimated low-frequency k-space data in MRI

PURPOSE

Single image super-resolution (SR) is highly desired in many fields but obtaining it is often technically limited in practice. The purpose of this study was to propose a simple, rapid and robust single image SR method in magnetic resonance (MR) imaging (MRI).

METHODS

The idea is based on the mathematical formulation of the intrinsic link in k-space between a given (modulus) low-resolution (LR) image and the desired SR image. The method consists of two steps: 1) estimating the low-frequency k-space data of the desired SR image from a single LR image; 2) reconstructing the SR image using the estimated low-frequency and zero-filled high-frequency k-space data. The method was evaluated on digital phantom images, physical phantom MR images and real brain MR images, and compared with existing SR methods.

RESULTS

The proposed SR method exhibited a good robustness by reaching a clearly higher PSNR (25.77dB) and SSIM (0.991) averaged over different noise levels in comparison with existing edge-guided nonlinear interpolation (EGNI) (PSNR=23.78dB, SSIM=0.983), zero-filling (ZF) (PSNR=24.09dB, SSIM=0.985) and total variation (TV) (PSNR=24.54dB, SSIM=0.987) methods while presenting the same order of computation time as the ZF method but being much faster than the EGNI or TV method. The average PSNR or SSIM over different slice images of the proposed method (PSNR=26.33 dB or SSIM=0.955) was also higher than the EGNI (PSNR=25.07dB or SSIM=0.952), ZF (PSNR=24.97dB or SSIM=0.950) and TV (PSNR=25.70dB or SSIM=0.953) methods, demonstrating its good robustness to variation in anatomical structure of the images. Meanwhile, the proposed method always produced less ringing artifacts than the ZF method, gave a clearer image than the EGNI method, and did not exhibit any blocking effect presented in the TV method. In addition, the proposed method yielded the highest spatial consistency in the inter-slice dimension among the four methods.

CONCLUSIONS

This study proposed a fast, robust and efficient single image SR method with high spatial consistency in the inter-slice dimension for clinical MR images by estimating the low-frequency k-space data of the desired SR image from a single spatial modulus LR image.

Super-Resolution Person Re-Identification With Semi-Coupled Low-Rank Discriminant Dictionary Learning

Person re-identification has been widely studied due to its importance in surveillance and forensics applications. In practice, gallery images are high resolution (HR), while probe images are usually low resolution (LR) in the identification scenarios with large variation of illumination, weather, or quality of cameras. Person re-identification in this kind of scenarios, which we call super-resolution (SR) person re-identification, has not been well studied. In this paper, we propose a semi-coupled low-rank discriminant dictionary learning (SLD²L) approach for SR person re-identification task. With the HR and LR dictionary pair and mapping matrices learned from the features of HR and LR training images, SLD²L can convert the features of the LR probe images into HR features. To ensure that the converted features have favorable discriminative capability and the learned dictionaries can well characterize intrinsic feature spaces of the HR and LR images, we design a discriminant term and a low-rank regularization term for SLD²L. Moreover, considering that low resolution results in different degrees of loss for different types of visual appearance features, we propose a multi-view SLD²L (MVSLD²L) approach, which can learn the type-specific dictionary pair and mappings for each type of feature. Experimental results on multiple publicly available data sets demonstrate the effectiveness of our proposed approaches for the SR person re-identification task.

Single Image Super-Resolution Using Global Regression Based on Multiple Local Linear Mappings

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.

Weighted Joint Sparse Representation for Removing Mixed Noise in Image

Joint sparse representation (JSR) has shown great potential in various image processing and computer vision tasks. Nevertheless, the conventional JSR is fragile to outliers. In this paper, we propose a weighted JSR (WJSR) model to simultaneously encode a set of data samples that are drawn from the same subspace but corrupted with noise and outliers. Our model is desirable to exploit the common information shared by these data samples while reducing the influence of outliers. To solve the WJSR model, we further introduce a greedy algorithm called weighted simultaneous orthogonal matching pursuit to efficiently approximate the global optimal solution. Then, we apply the WJSR for mixed noise removal by jointly coding the grouped nonlocal similar image patches. The denoising performance is further improved by incorporating it with the global prior and the sparse errors into a unified framework. Experimental results show that our denoising method is superior to several state-of-the-art mixed noise removal methods.

Pairwise Operator Learning for Patch-Based Single-Image Super-Resolution

Motivated by the fact that image patches could be inherently represented by matrices, single-image super-resolution is treated as a problem of learning regression operators in a matrix space in this paper. The regression operators that map low-resolution image patches to high-resolution image patches are generally defined by the left and right multiplication operators. The pairwise operators are, respectively, used to extract the raw and column information of low-resolution image patches for recovering high-resolution estimations. The patch-based regression algorithm possesses three favorable properties. First, the proposed super-resolution algorithm is efficient during both training and testing, because image patches are treated as matrices. Second, the data storage requirement of the optimal pairwise operator is far less than most popular single-image super-resolution algorithms, because only two small sized matrices need to be stored. Last, the super-resolution performance is competitive with most popular single-image super-resolution algorithms, because both raw and column information of image patches is considered. Experimental results show the efficiency and effectiveness of the proposed patch-based single-image super-resolution algorithm.

Adaptive Subspace-Based Inverse Projections via Division Into Multiple Sub-Problems for Missing Image Data Restoration

This paper presents adaptive subspace-based inverse projections via division into multiple sub-problems (ASIP-DIMSs) for missing image data restoration. In the proposed method, a target problem for estimating missing image data is divided into multiple sub-problems, and each sub-problem is iteratively solved with the constraints of other known image data. By projection into a subspace model of image patches, the solution of each sub-problem is calculated, where we call this procedure "subspace-based inverse projection" for simplicity. The proposed method can use higher dimensional subspaces for finding unique solutions in each sub-problem, and successful restoration becomes feasible, since a high level of image representation performance can be preserved. This is the biggest contribution of this paper. Furthermore, the proposed method generates several subspaces from known training examples and enables derivation of a new criterion in the above framework to adaptively select the optimal subspace for each target patch. In this way, the proposed method realizes missing image data restoration using ASIP-DIMS. Since our method can estimate any kind of missing image data, its potential in two image restoration tasks, image inpainting and super-resolution, based on several methods for multivariate analysis is also shown in this paper.

Multiple Representations-Based Face Sketch-Photo Synthesis

Face sketch-photo synthesis plays an important role in law enforcement and digital entertainment. Most of the existing methods only use pixel intensities as the feature. Since face images can be described using features from multiple aspects, this paper presents a novel multiple representations-based face sketch-photo-synthesis method that adaptively combines multiple representations to represent an image patch. In particular, it combines multiple features from face images processed using multiple filters and deploys Markov networks to exploit the interacting relationships between the neighboring image patches. The proposed framework could be solved using an alternating optimization strategy and it normally converges in only five outer iterations in the experiments. Our experimental results on the Chinese University of Hong Kong (CUHK) face sketch database, celebrity photos, CUHK Face Sketch FERET Database, IIIT-D Viewed Sketch Database, and forensic sketches demonstrate the effectiveness of our method for face sketch-photo synthesis. In addition, cross-database and database-dependent style-synthesis evaluations demonstrate the generalizability of this novel method and suggest promising solutions for face identification in forensic science.

Reconstruction of 7T-Like Images From 3T MRI

In the recent MRI scanning, ultra-high-field (7T) MR imaging provides higher resolution and better tissue contrast compared to routine 3T MRI, which may help in more accurate and early brain diseases diagnosis. However, currently, 7T MRI scanners are more expensive and less available at clinical and research centers. These motivate us to propose a method for the reconstruction of images close to the quality of 7T MRI, called 7T-like images, from 3T MRI, to improve the quality in terms of resolution and contrast. By doing so, the post-processing tasks, such as tissue segmentation, can be done more accurately and brain tissues details can be seen with higher resolution and contrast. To do this, we have acquired a unique dataset which includes paired 3T and 7T images scanned from same subjects, and then propose a hierarchical reconstruction based on group sparsity in a novel multi-level Canonical Correlation Analysis (CCA) space, to improve the quality of 3T MR image to be 7T-like MRI. First, overlapping patches are extracted from the input 3T MR image. Then, by extracting the most similar patches from all the aligned 3T and 7T images in the training set, the paired 3T and 7T dictionaries are constructed for each patch. It is worth noting that, for the training, we use pairs of 3T and 7T MR images from each training subject. Then, we propose multi-level CCA to map the paired 3T and 7T patch sets to a common space to increase their correlations. In such space, each input 3T MRI patch is sparsely represented by the 3T dictionary and then the obtained sparse coefficients are used together with the corresponding 7T dictionary to reconstruct the 7T-like patch. Also, to have the structural consistency between adjacent patches, the group sparsity is employed. This reconstruction is performed with changing patch sizes in a hierarchical framework. Experiments have been done using 13 subjects with both 3T and 7T MR images. The results show that our method outperforms previous methods and is able to recover better structural details. Also, to place our proposed method in a medical application context, we evaluated the influence of post-processing methods such as brain tissue segmentation on the reconstructed 7T-like MR images. Results show that our 7T-like images lead to higher accuracy in segmentation of white matter (WM), gray matter (GM), cerebrospinal fluid (CSF), and skull, compared to segmentation of 3T MR images.

SERF: A Simple, Effective, Robust, and Fast Image Super-Resolver From Cascaded Linear Regression

Example learning-based image super-resolution techniques estimate a high-resolution image from a low-resolution input image by relying on high- and low-resolution image pairs. An important issue for these techniques is how to model the relationship between high- and low-resolution image patches: most existing complex models either generalize hard to diverse natural images or require a lot of time for model training, while simple models have limited representation capability. In this paper, we propose a simple, effective, robust, and fast (SERF) image super-resolver for image super-resolution. The proposed super-resolver is based on a series of linear least squares functions, namely, cascaded linear regression. It has few parameters to control the model and is thus able to robustly adapt to different image data sets and experimental settings. The linear least square functions lead to closed form solutions and therefore achieve computationally efficient implementations. To effectively decrease these gaps, we group image patches into clusters via k-means algorithm and learn a linear regressor for each cluster at each iteration. The cascaded learning process gradually decreases the gap of high-frequency detail between the estimated high-resolution image patch and the ground truth image patch and simultaneously obtains the linear regression parameters. Experimental results show that the proposed method achieves superior performance with lower time consumption than the state-of-the-art methods.

Antipodally Invariant Metrics for Fast Regression-Based Super-Resolution

Dictionary-based super-resolution (SR) algorithms usually select dictionary atoms based on the distance or similarity metrics. Although the optimal selection of the nearest neighbors is of central importance for such methods, the impact of using proper metrics for SR has been overlooked in literature, mainly due to the vast usage of Euclidean distance. In this paper, we present a very fast regression-based algorithm, which builds on the densely populated anchored neighborhoods and sublinear search structures. We perform a study of the nature of the features commonly used for SR, observing that those features usually lie in the unitary hypersphere, where every point has a diametrically opposite one, i.e., its antipode, with same module and angle, but the opposite direction. Even though, we validate the benefits of using antipodally invariant metrics, most of the binary splits use Euclidean distance, which does not handle antipodes optimally. In order to benefit from both the worlds, we propose a simple yet effective antipodally invariant transform that can be easily included in the Euclidean distance calculation. We modify the original spherical hashing algorithm with this metric in our antipodally invariant spherical hashing scheme, obtaining the same performance as a pure antipodally invariant metric. We round up our contributions with a novel feature transform that obtains a better coarse approximation of the input image thanks to iterative backprojection. The performance of our method, which we named antipodally invariant SR, improves quality (Peak Signal to Noise Ratio) and it is faster than any other state-of-the-art method.

Single Image Super-Resolution Based on Sparse Representation with Adaptive Dictionary Selection

Sparse representation theory has attracted much attention, and has been successfully used in image super-resolution (SR) reconstruction. However, it could only provide the local prior of image patches. Field of experts (FoE) is a way to develop the generic and expressive prior of the whole image. The algorithm proposed in this paper uses the FoE model as the global constraint of SR reconstruction problem to pre-process the low-resolution image. Since a single dictionary could not accurately represent different types of image patches, our algorithm classifies the sample patches composed of pre-processed image and high-resolution image, obtains the sub-dictionaries by training, and adaptively selects the most appropriate sub-dictionary for reconstruction according to the pyramid histogram of oriented gradients feature of image patches. Furthermore, in order to reduce the computational complexity, our algorithm makes use of edge detection, and only applies SR reconstruction based on sparse representation to the edge patches of the test image. Nonedge patches are directly replaced by the pre-processing results of FoE model. Experimental results show that our algorithm can effectively guarantee the quality of the reconstructed image, and reduce the computation time to a certain extent.

Joint Prior Learning for Visual Sensor Network Noisy Image Super-Resolution

The visual sensor network (VSN), a new type of wireless sensor network composed of low-cost wireless camera nodes, is being applied for numerous complex visual analyses in wild environments, such as visual surveillance, object recognition, etc. However, the captured images/videos are often low resolution with noise. Such visual data cannot be directly delivered to the advanced visual analysis. In this paper, we propose a joint-prior image super-resolution (JPISR) method using expectation maximization (EM) algorithm to improve VSN image quality. Unlike conventional methods that only focus on upscaling images, JPISR alternatively solves upscaling mapping and denoising in the E-step and M-step. To meet the requirement of the M-step, we introduce a novel non-local group-sparsity image filtering method to learn the explicit prior and induce the geometric duality between images to learn the implicit prior. The EM algorithm inherently combines the explicit prior and implicit prior by joint learning. Moreover, JPISR does not rely on large external datasets for training, which is much more practical in a VSN. Extensive experiments show that JPISR outperforms five state-of-the-art methods in terms of both PSNR, SSIM and visual perception.

Single-Image Super-Resolution Using Active-Sampling Gaussian Process Regression

As well known, Gaussian process regression (GPR) has been successfully applied to example learning-based image super-resolution (SR). Despite its effectiveness, the applicability of a GPR model is limited by its remarkably computational cost when a large number of examples are available to a learning task. For this purpose, we alleviate this problem of the GPR-based SR and propose a novel example learning-based SR method, called active-sampling GPR (AGPR). The newly proposed approach employs an active learning strategy to heuristically select more informative samples for training the regression parameters of the GPR model, which shows significant improvement on computational efficiency while keeping higher quality of reconstructed image. Finally, we suggest an accelerating scheme to further reduce the time complexity of the proposed AGPR-based SR by using a pre-learned projection matrix. We objectively and subjectively demonstrate that the proposed method is superior to other competitors for producing much sharper edges and finer details.