Low-rank representation (LRR) is an effective method for subspace clustering and has found wide applications in computer vision machine learning. The existing LRR solver based on the alternating direction (ADM). It suffers from $O(n^3)$ computation complexity due to matrix-matrix multiplications matrix inversions, even if partial SVD used. Moreover, introducing auxiliary variables also slows down convergence. Such a heavy load prevents large scale applications. In this paper, we generalize...

We propose a simple yet effective L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> -regularized prior based on intensity and gradient for text image deblurring. The proposed is motivated by observing distinct properties of images. Based this prior, we develop an efficient optimization method to generate reliable intermediate results kernel estimation. does not require any complex filtering strategies select salient edges which are...

We propose a simple yet effective L0-regularized prior based on intensity and gradient for text image deblurring. The proposed is distinctive properties of images, with which we develop an efficient optimization algorithm to generate reliable intermediate results kernel estimation. does not require any heuristic edge selection methods, are critical the state-of-the-art edge-based deblurring methods. discuss relationship other methods present how select salient edges more principally. For...

Higher-order low-rank tensors naturally arise in many applications including hyperspectral data recovery, video inpainting, seismic recon- struction, and so on. We propose a new model to recover tensor by simultaneously performing matrix factorizations the all-mode ma- tricizations of underlying tensor. An alternating minimization algorithm is applied solve model, along with two adaptive rank-adjusting strategies when exact rank not known. Phase transition plots reveal that our can variety...

Benefiting from its effectiveness in subspace segmentation, low-rank representation (LRR) and variations have many applications computer vision pattern recognition, such as motion image saliency detection, semisupervised learning. It is known that the standard LRR can only work well under assumption all subspaces are independent. However, this cannot be guaranteed real-world problems. This paper addresses problem provides an extension of LRR, named structure-constrained (SC-LRR), to analyze...

Low contrast and poor quality are main problems in the production of medical images. By using wavelet transform Haar transform, a novel image enhancement approach is proposed. First, was decomposed with transform. Secondly, all high-frequency sub-images were Thirdly, noise frequency field reduced by soft-threshold method. Fourthly, coefficients enhanced different weight values sub-images. Then, obtained through inverse Lastly, image's histogram stretched nonlinear equalisation. Experiments...

Deep convolutional neural networks (CNNs) have contributed to the significant progress of single-image super-resolution (SISR) field. However, majority existing CNN-based models maintain high performance with massive parameters and exceedingly deeper structures. Moreover, several algorithms essentially underused low-level features, thus causing relatively low performance. In this article, we address these problems by exploring two strategies based on novel local wider residual blocks (LWRBs)...

Image dehazing is a well-known ill-posed problem, which usually requires some image priors to make the problem well-posed. We propose an effective iteration algorithm with deep CNNs learn haze-relevant for dehazing. formulate as minimization of variational model favorable data fidelity terms and prior regularize model. solve based on classical gradient descent method built-in so that iteration-wise atmospheric light, transmission map clear can be well estimated. Our combines properties both...

Estimating blur kernels from real world images is a challenging problem as the linear image formation assumption does not hold when significant outliers, such saturated pixels and non-Gaussian noise, are present. While some existing non-blind deblurring algorithms can deal with outliers to certain extent, few blind methods developed well estimate blurred outliers. In this paper, we present an algorithm address by exploiting reliable edges removing in intermediate latent images, thereby...

Object motion blur is a challenging problem as the foreground and background in scenes undergo different types of image degradation due to movements various directions speed. Most object deblurring methods address this by segmenting blurred images into regions where kernels are estimated applied for restoration. Segmentation on difficult ambiguous pixels between regions, but it plays an important role deblurring. To these problems, we propose novel model The proposed developed based maximum...

Deblurring images with outliers has attracted considerable attention recently. However, existing algorithms usually involve complex operations which increase the difficulty of blur kernel estimation. In this paper, we propose a simple yet effective blind image deblurring algorithm to handle blurred outliers. The proposed method is motivated by observation that in significantly affect goodness-of-fit function approximation. Therefore, an model data fidelity term so have little effect on does...

Rain removal is an important but challenging computer vision task as rain streaks can severely degrade the visibility of images that may make other visions or multimedia tasks fail to work. Previous works mainly focused on feature extraction and processing neural network structure, while current methods already achieve remarkable results, training based single structure without considering cross-scale relationship cause information drop-out. In this paper, we explore manner between networks...

New relativistic atomic ground-state wave functions calculated by using the Dirac–Fock program package GRASP92 [Parpia et al. (1996). Comput. Phys. Commun. 94, 249–271; Su & Coppens (1997). Acta Cryst. A53, 749–762, (1998). A54, 357] for atoms H through Kr (Z = 1–36) have been fitted a linear combination of Slater-type nonlinear least-squares procedure. These analytical expressions allow derivation closed-form X-ray scattering factors core and valence electrons densities electrostatic...

Subspace clustering and feature extraction are two of the most commonly used unsupervised learning techniques in computer vision pattern recognition. State-of-the-art for subspace make use recent advances sparsity rank minimization. However, existing computationally expensive may result degenerate solutions that degrade performance case insufficient data sampling. To partially solve these problems, inspired by work on matrix factorization, this paper proposes fixed-rank representation (FRR)...

Blind image deblurring is a challenging problem in computer vision and processing. In this paper, we propose new <i xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">l</i> <sup xmlns:xlink="http://www.w3.org/1999/xlink">0</sup> -regularized approach to estimate blur kernel from single blurred by regularizing the sparsity property of natural images. Furthermore, introducing an adaptive structure map process, our method able restore useful salient edges...

In the field of multimedia, single image deraining is a basic pre-processing work, which can greatly improve visual effect subsequent high-level tasks in rainy conditions. this paper, we propose an effective algorithm, called JDNet, to solve problem and conduct segmentation detection task for applications. Specifically, considering important information on multi-scale features, Scale-Aggregation module learn features with different scales. Simultaneously, Self-Attention introduced match or...

Subspace clustering and feature extraction are two of the most commonly used unsupervised learning techniques in computer vision pattern recognition. State-of-the-art for subspace make use recent advances sparsity rank minimization. However, existing computationally expensive may result degenerate solutions that degrade performance case insufficient data sampling. To partially solve these problems, inspired by work on matrix factorization, this paper proposes fixed-rank representation (FRR)...

We present a simple yet effective unpaired learning based image rain removal method from an set of synthetic images and real rainy by exploring the properties maps. The proposed algorithm mainly consists semi-supervised part knowledge distillation part. estimates map reconstructs derained on well-established layer separation principle. To facilitate removal, we develop direction regularizer to constrain estimation network in With estimated maps part, first synthesize new paired adding...

A fermion-spin transformation is used to implement the charge-spin separation, and developed study low-dimensional t-J model. In this approach, charge spin degrees of freedom physical electron are separated, degree represented by a spinless fermion while representd hard-core boson. The on-site local constraint for single occupancy satisfied even in mean-field approximation sum rule obeyed. This approach can be applied both one two-dimensional systems. one-dimensional case, spinon as well...

We have investigated the physical effects of Dzyaloshinskii-Moriya (DM) interaction in copper benzoate. In low-field limit, spin gap is found to vary as H(2/3)ln((1/6)(J/mu(B)H(s)) (H(s): an effective staggered field induced by external H) agreement with prediction conformal theory, while magnetization varies H(1/3) and ln((1/3)(J/mu(B)H(s)) correction predicted theory not confirmed. The linear scaling relation between momentum shift broken. determined coupling constant DM given a complete...

Many problems in machine learning and other fields can be (re)for-mulated as linearly constrained separable convex programs. In most of the cases, there are multiple blocks variables. However, traditional alternating direction method (ADM) its linearized version (LADM, obtained by linearizing quadratic penalty term) for two-block case cannot naively generalized to solve multi-block case. So is great demand on extending ADM based methods this paper, we propose LADM with parallel splitting...