Nonnegative matrix factorization with Hessian regularizer

Because of the underlying data structure preserved by the manifold regularization term, the Nonnegative matrix factorization (NMF) with manifold regularizer demonstrates an advantage over the variants of NMF for many data analysis tasks. Currently, the Laplacian regularizer is commonly used as the smooth operator to preserve the locality of data space. However, with the Laplacian regularizer, coding vectors are biased to a constant, which leads to a lack of extrapolating power. Thus, the locality of data space cannot be preserved, as would be expected. To address this drawback, a novel variant of NMF, namely HsNMF, is proposed, where the Hessian regularization term is incorporated into the traditional NMF framework. Because Hessian Energy favors the functions whose values vary linearly with respect to the geodesics of the data manifold, the local structure of data space is more effectively preserved. Clustering and classification experimental results on real-world image datasets demonstrate that our proposed NMF is superior to the variants of NMF based on Laplacian Embedding.     

去除乘性噪声的重加权各向异性全变差模型

恢复含乘性噪声的图像是当前图像处理的重要研究课题. 本文提出基于迭代重加权的各向异性全变差(Total variation, TV)模型. 新模型中, 假定乘性噪声服从Gamma分布. 正则项采用加权的各向异性全变差, 其中, 自适应权函数由期望最大(Expectation maximization, EM)算法得到. 新模型在有效去噪的同时, 较好地保留了图像的边缘和细节信息, 同时能够有效地抑制"阶梯效应". 数值实验验证了新模型的效果.     

A fast minimization method for blur and multiplicative noise removal

Multiplicative noise and blur removal problems have attracted much attention in recent years. In this paper, we propose an efficient minimization method to recover images from input blurred and multiplicative noisy images. In the proposed algorithm, we make use of the logarithm to transform blurring and multiplicative noise problems into additive image degradation problems, and then employ l ₁-norm to measure in the data-fitting term and the total variation to measure the regularization term. The alternating direction method of multipliers (ADMM) is used to solve the corresponding minimization problem. In order to guarantee the convergence of the ADMM algorithm, we approximate the associated nonconvex domain of the minimization problem by a convex domain. Experimental results are given to demonstrate that the proposed algorithm performs better than the other existing methods in terms of speed and peak signal noise ratio.     

A well-balanced and adaptive variational model for the removal of mixed noise

Image denoising is one of the fundamental problems concerning image processing. Over the last decade mathematical models based on partial differential equations and variational techniques have led to superior results related to denoising problems. The additive noise models have been studied extensively, however, the reconstruction of images corrupted by nonadditive noise has not yet been thoroughly studied. In this paper, a novel variational method for the reconstruction of images corrupted by non-uniformly distributed noise is presented. The proposed model includes a balance between the data term and the regularization term in the energy functional, which takes into account the statistical control of the parameters and the position of the noisy points related to the edges presented in the image. The parameters are determined by the given initial noisy image. The obtained results have shown the effectiveness and robustness of the proposed model and in restoring images with multiplicative noise or mixed Gaussian noise, while preserving edges and small structures belonging to the image.     

A new fast multiphase image segmentation algorithm based on nonconvex regularizer

We present a new variational model for the soft multiphase image segmentation. In the model, we introduce a nonconvex regularizer on the membership functions which are used as indicators of different homogeneous regions. The nonconvex regularizer performs better than the usual convex ones in that (i) it well preserves geometric shapes of the homogeneous regions, and (ii) it protects edges from oversmoothing which is a common drawback of the convex regularizer. To solve the nonconvex minimization problem, we design a new fast alternative iteration algorithm, which is robust to the setting of the parameters in the model. We conduct comprehensive experiments to measure the performance of the algorithm in terms of visual evaluation and a variety of quantitative indices for image segmentation. The algorithm achieves more accurate results compared to other well-known convex variational methods for image segmentation.     

On Objective Function, Regularizer, and Prediction Error of a Learning Algorithm for Dealing With Multiplicative Weight Noise

In this paper, an objective function for training a functional link network to tolerate multiplicative weight noise is presented. Basically, the objective function is similar in form to other regularizer-based functions that consist of a mean square training error term and a regularizer term. Our study shows that under some mild conditions the derived regularizer is essentially the same as a weight decay regularizer. This explains why applying weight decay can also improve the fault-tolerant ability of a radial basis function (RBF) with multiplicative weight noise. In accordance with the objective function, a simple learning algorithm for a functional link network with multiplicative weight noise is derived. Finally, the mean prediction error of the trained network is analyzed. Simulated experiments on two artificial data sets and a real-world application are performed to verify theoretical result.     

Total variation and high-order total variation adaptive model for restoring blurred images with Cauchy noise

In this paper, we propose a novel model to restore an image corrupted by blur and Cauchy noise. The model is composed of a data fidelity term and two regularization terms including total variation and high-order total variation. Total variation provides well-preserved edge features, but suffers from staircase effects in smooth regions, whereas high-order total variation can alleviate staircase effects. Moreover, we introduce a strategy for adaptively selecting regularization parameters. We develop an efficient alternating minimization algorithm for solving the proposed model. Numerical examples suggest that the proposed method has the advantages of better preserving edges and reducing staircase effects.     

Multiplicative Denoising Based on Linearized Alternating Direction Method Using Discrepancy Function Constraint

The multiplicative noise (speckle) in coherent imaging systems such as synthetic aperture radar makes it difficult to interpret observed images. Recently, the total variation (TV) models have received much interest in removing the speckle due to the strong edge preserving ability and low computational cost of the TV regularizer. However, the classical methods have difficulties in two aspects: one is how to efficiently compute the solution of the models with special data-fidelity terms, the other is how to choose the regularization parameter since the variational models are rather sensitive to the parameter. In this paper, we propose a new linearized alternating direction method, which is able to handle the data-fidelity term efficiently, and meanwhile estimate the optimal value of the regularization parameter exactly based on a discrepancy function constraint. We further establish the global convergence of the proposed algorithm. Numerical experiments demonstrate that our methods overall outperform the current state-of-the-art methods for multiplicative noise removal.     

Multiplicative noise removal via combining total variation and wavelet frame

Total variation (TV) regularization has been proved effective for cartoon images restoration however it produces staircase effects, and properly wavelet frames were confirmed to provide a more smoothing approximation to the original image. In this paper, a new model for multiplicative noise removal was proposed, which combines wavelet frame-based regularization and TV regularization. A modified proximal linearized alternating direction method is developed to solve the proposed model, considering that adding a new regularization term to the TV model would yield more parameters, which will result in computational difficulties. For the new model, the existence of solution and the convergence property of the proposed algorithm are proved. Numerical experiments have proved that the proposed model has a superior performance in terms of the peak signal-to-noise ratio and the relative error values for non-piecewise constant images when compared with some state-of-the-art multiplicative noise removal models.     

Image Deblurring in the Presence of Impulsive Noise

Consider the problem of image deblurring in the presence of impulsive noise. Standard image deconvolution methods rely on the Gaussian noise model and do not perform well with impulsive noise. The main challenge is to deblur the image, recover its discontinuities and at the same time remove the impulse noise. Median-based approaches are inadequate, because at high noise levels they induce nonlinear distortion that hampers the deblurring process. Distinguishing outliers from edge elements is difficult in current gradient-based edge-preserving restoration methods. The suggested approach integrates and extends the robust statistics, line process (half quadratic) and anisotropic diffusion points of view. We present a unified variational approach to image deblurring and impulse noise removal. The objective functional consists of a fidelity term and a regularizer. Data fidelity is quantified using the robust modified L ¹ norm, and elements from the Mumford-Shah functional are used for regularization. We show that the Mumford-Shah regularizer can be viewed as an extended line process. It reflects spatial organization properties of the image edges, that do not appear in the common line process or anisotropic diffusion. This allows to distinguish outliers from edges and leads to superior experimental results.     

Joint sparsity and fidelity regularization for segmentation-driven CT image preprocessing基于稀疏性和保真性正则化模型的分割驱动的CT图像预处理研究

In this paper, we propose a novel segmentation-driven computed tomography (CT) image preprocessing approach. The proposed approach, namely, joint sparsity and fidelity regularization (JSFR) model can be regarded as a generalized total variation (TV) denoising model or a generalized sparse representation denoising model by adding an additional gradient fidelity regularizer and a stronger gradient sparsity regularizer. Thus, JSFR model consists of three terms: intensity fidelity term, gradient fidelity term, and gradient sparsity term. The interactions and counterbalance of these terms make JSFR model has the ability to reduce intensity inhomogeneities and improve edge ambiguities of a given image. Experimental results carried out on the real dental cone-beam CT data demonstrate the effectiveness and usefulness of JSFR model for CT image intensity homogenization, edge enhancement, as well as tissue segmentation.     

Multiplicative noise removal via adaptive learned dictionaries and TV regularization

Multiplicative noise removal is a key issue in image processing problem. While a large amount of literature on this subject are total variation (TV)-based and wavelet-based methods, recently sparse representation of images has shown to be efficient approach for image restoration. TV regularization is efficient to restore cartoon images while dictionaries are well adapted to textures and some tricky structures. Following this idea, in this paper, we propose an approach that combines the advantages of sparse representation over dictionary learning and TV regularization method. The method is proposed to solve multiplicative noise removal problem by minimizing the energy functional, which is composed of the data-fidelity term, a sparse representation prior over adaptive learned dictionaries, and TV regularization term. The optimization problem can be efficiently solved by the split Bregman algorithm. Experimental results validate that the proposed model has a superior performance than many recent methods, in terms of peak signal-to-noise ratio, mean absolute-deviation error, mean structure similarity, and subjective visual quality.     

优化加权TV的复合正则化压缩感知图像重建

目的：压缩感知理论突破了传统的Shanon-Nyquist采样定理的限制,能够以较少的采样值来进行原信号的恢复。针对压缩感知图像重建问题,本文提出了一种基于优化加权全变差（Total Variation, TV）的复合正则化压缩感知图像重建模型。方法：提出的重建模型是以TV正则化模型为基础的。首先,为克服传统TV正则化会导致重建图像的边缘和纹理细节部分模糊或丢失的缺点,本文引入图像的梯度信息估计权重,构建加权TV的重建模型。其次,利用全变差去噪（Rudin–Osher–Fatemi,ROF）模型对权重进行优化估计,从而减少计算权重时受噪声的影响。再次,本文将非局部结构相似性先验和局部自回归性先验引入提出的加权TV模型,得到优化加权TV的复合正则化重建模型。最后,结合投影法和算子分裂法对优化模型求解。结果：针对自然图像的不同特性,本文使用复合正则化先验进行建模,实验表明上述重建问题通过我们的方法得到了很好的解决,加权TV正则化先验使得图像的平坦区域和强边重建较好,而非局部结构相似性先验和局部自回归性先验能够保证图像的精细结构部分的重建效果。结论：本文提出了一种新的复合正则化压缩感知重建模型。与其它基于TV正则化的重建模型相比,实验结果表明本文模型的重建性能无论是在视觉效果还是在客观评价指标上都有明显的提高。     

Fast multi-channel image reconstruction using a novel two-dimensional algorithm

Recently, two 2D algorithms for super resolution image reconstruction based on a matrix observation model were presented. They can greatly reduce computational cost and storage requirement but are suitable for the cases of face images or no warping operator. In this paper, for wide applications we propose a novel 2D algorithm to reconstruct a high-resolution image from multiple warped and degraded low-resolution images. The proposed 2D algorithm minimizes a new cost function with two regularization terms where one is the Laplacian regularization term for robustness to noise and another is learning term for more high frequency information. Simulation results show that the proposed 2D algorithm can obtain better results in terms of both PSNR and visual quality than the two existing 2D algorithms.     

一种自适应鲁棒最小体积高光谱解混算法

对高光谱图像解混的目的在于从低空间分辨率的高光谱图像中找到端元与对应的丰度.本文根据解混算法中的最小体积准则,提出了一种自适应鲁棒最小体积高光谱解混算法（Robust minimum volume based algorithm with automatically estimating regularization parameters for hyperspectral unmixing,RMVHU）.本算法通过引入负数惩罚正则项,替换了同类算法中的丰度非负性约束（Non-negativity constraint,ANC）,使算法对图像中的噪声与异常值具有更强的鲁棒性;采用循环最小化方法,将非凸优化问题分解为凸优化子问题,然后应用交替方向乘子法解决随着像素点个数增大带来的求解困难问题;对于正则项系数,本算法提出了一种自适应调整策略,提高了算法的收敛性,并且通过定性分析,说明了该调整方法的合理性.将算法应用于合成数据与实际数据,实验结果表明,与同类算法相比,本文提出的算法能够取得更为优秀的效果.     

A total variation recursive space-variant filter for image denoising

Total Variation (TV) regularization is a widely used convex but non-smooth regularizer in image restoration and reconstruction. Many algorithms involve solving a denoising problem as an intermediate step or in each iteration. Most existing solvers were proposed in the context of a specific application. In this paper, we propose a denoising method which can be used as a proximal mapping (denoising operator) for noises other than additive and Gaussian. We formulate the Maximum A-Posteriori (MAP) estimation in terms of a spatially adaptive and recursive filtering operation on the Maximum Likelihood (ML) estimate. The only dependence on the model is the ML estimate and the second order derivative, which are computed at the beginning and remain fixed throughout the iterative process. The proposed method generalizes the MAP estimation with a quadratic regularizer using an infinite impulse response filter, to the case with TV regularization. Due to the fact that TV is non-smooth and has spatial dependencies, the resulting filter after reweighted least squares formulation of the TV term, is recursive and spatially variant. The proposed method is an instance of the Majorization–Minimization (MM) algorithms, for which convergence conditions are defined and can be shown to be satisfied by the proposed method. The method can also be extended to image inpainting and higher order TV in an intuitively straight-forward manner.     

Multiregion level-set partitioning of synthetic aperture radar images

The purpose of this study is to investigate synthetic aperture radar (SAR) image segmentation into a given but arbitrary number of gamma homogeneous regions via active contours and level sets. The segmentation of SAR images is a difficult problem due to the presence of speckle which can be modeled as strong, multiplicative noise. The proposed algorithm consists of evolving simple closed planar curves within an explicit correspondence between the interiors of curves and regions of segmentation to minimize a criterion containing a term of conformity of data to a speckle model of noise and a term of regularization. Results are shown on both synthetic and real images.     

自适应非局部数据保真项和双边总变分的图像去噪模型

针对常见去噪方法容易造成特定区域过度平滑、奇异结构残余噪声以及产生阶梯效应和对比度损失等问题,提出一种自适应非局部数据保真项和双边总变分的图像去噪模型,建立了自适应非局部正则化能量泛函和相应的变分框架。首先,对噪声图像利用自适应权值的非局部均值求得数据拟合项;其次,引入双边总变分正则化项,利用正则化系数来适度平衡数据拟合项和正则化项的影响;最后,通过能量函数最小化对不同的噪声统计快速求得最优解,从而达到降低残余噪声并纠正过度平滑的目的。通过理论分析和针对模拟噪声图像与真实噪声图像的实验结果表明,所提出的图像去噪模型能够较好地处理具有不同统计特性的图像噪声,与自适应非局部均值滤波去噪相比,所提算法的峰值信噪比（PSNR）值最多可以得到0.6 dB的改善;与全变分正则化图像去噪算法比较,所提算法的主观视觉效果明显更好,在去噪的同时图像纹理和边缘等细节信息保护得更好,PSNR值最多可以提高10 dB,而多尺度结构相似性度（MS-SSIM）指标可以提升0.3。因此,所提出的图像去噪模型可以在理论上更好地探讨如何合理处理噪声和图像内容本身的高频细节信息,在视频和图像分辨率提升等领域也具有良好的实际应用价值。     

Adaptive active contour model driven by image data field for image segmentation with flexible initialization

In this paper, a novel adaptive active contour model based on image data field for image segmentation with robust and flexible initializations is proposed. We firstly construct a new external energy term deduced from the image data field that drives the level set function to move in the opposite direction along the boundaries of object and an adaptive length regularization term based on the image local entropy. The designed external energy and length regularization term are then incorporated into a variationlevel set framework with an additional penalizing energy term. Due to the adaptive sign–changing property of the external energy and the adaptive length regularization term, the proposed model can tackle images with clutter background and noise, the level set function can be initialized as any bounded functions (e.g., constant function), which implies the proposed model is robust to initialization of contours. Experimental results on both synthetic and real images from different modalities confirm the effectiveness and competivive performance of the proposed method compared with other representative models.

     

Ultrasound image restoration based on a learned dictionary and a higher-order MRF

The restoration of images degraded by blur and multiplicative noise is a critical preprocessing step in medical ultrasound images which exhibit clinical diagnostic features of interest. This paper proposes a novel non-smooth non-convex variational model for ultrasound images denoising and deblurring motivated by the successes of sparse representation of images and FoE based approaches. Dictionaries are well adapted to textures and extended to arbitrary image sizes by defining a global image prior, while FoE image prior explicitly characterizes the statistics properties of natural image. Following these ideas, the new model is composed of the data-fidelity term, the sparse and redundant representations via learned dictionaries, and the FoE image prior model. The iPiano algorithm can efficiently deal with this optimization problem. The new proposed model is applied to several simulated images and real ultrasound images. The experimental results of denoising and deblurring show that proposed method gives a better visual effect by efficiently removing noise and preserving details well compared with two state-of-the-art methods.