Efficient Dynamic Parallel MRI Reconstruction for the Low-Rank Plus Sparse Model.

The low-rank plus sparse (L+S) decomposition model enables the reconstruction of undersampled dynamic magnetic resonance imaging (MRI) data. Solving for the L and S components is anonsmooth composite convex minimization problem. While current techniques for this model are based on the classical iterative soft thresholding algorithm (ISTA), accelerated methods can be applied to obtain faster rate of convergence of the algorithm. This paper proposes two alternative methods for solving the L+S problem, one based on the fast iterative shrinkage-thresholding algorithm (FISTA), and the other based on the recent proximal optimized gradient method (POGM). Numerical results suggest faster convergence than the traditional ISTA, while preserving its computational simplicity.

Image reconstruction for Electrical Resistance Tomography (ERT) is an ill-posed nonlinear inverse problem. Considering the influence of the sparse measurement data on the quality of the reconstructed image, the l1-regularized least-squares program (l1 regularized LSP) is introduced to solve the inverse problem in this paper. To meet the need of high speed in ERT, the fast iterative shrinkage-thresholding algorithm (FISTA) is employed for image reconstruction in our work. Simulation results of the FISTA and l1_ls algorithm show that the l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sub> regularized LSP is superior to the l2 regularization method, especially in avoiding the over-smoothing of the reconstructed image. In addition, to improve the convergence speed and imaging quality in FISTA algorithm, the initial guess is calculated with the conjugate gradient method. Comparative simulation results demonstrate the feasibility of FISTA in ERT system and its advantage over the l1_ls regularization method.

Multiplicative noise (also known as speckle noise) models are central to the study of coherent imaging systems, such as synthetic aperture radar and sonar, and ultrasound and laser imaging. These models introduce two additional layers of difficulties with respect to the standard Gaussian additive noise scenario: (1) the noise is multiplied by (rather than added to) the original image; (2) the noise is not Gaussian, with Rayleigh and Gamma being commonly used densities. These two features of multiplicative noise models preclude the direct application of most state-of-the-art algorithms, which are designed for solving unconstrained optimization problems where the objective has two terms: a quadratic data term (log-likelihood), reflecting the additive and Gaussian nature of the noise, plus a convex (possibly nonsmooth) regularizer (e.g., a total variation or wavelet-based regularizer/prior). In this paper, we address these difficulties by: (1) converting the multiplicative model into an additive one by taking logarithms, as proposed by some other authors; (2) using variable splitting to obtain an equivalent constrained problem; and (3) dealing with this optimization problem using the augmented Lagrangian framework. A set of experiments shows that the proposed method, which we name MIDAL (multiplicative image denoising by augmented Lagrangian), yields state-of-the-art results both in terms of speed and denoising performance.

In this paper, we consider a regularized least squares problem subject to convex constraints. Our algorithm is based on the superiorization technique, equipped with a new step size rule which uses subgradient projections. The superiorization method is a two-step method where one step reduces the value of the penalty term and the other step reduces the residual of the underlying linear system (using an algorithmic operator T). For the new step size rule, we present a convergence analysis for the case when T belongs to a large subclass of strictly quasi-nonexpansive operators. To examine our algorithm numerically, we consider box constraints and use the total variation (TV) functional as a regularization term. The specific test cases are chosen from computed tomography using both noisy and noiseless data. We compare our algorithm with previously used parameters in superiorization. The T operator is based on sequential block iteration (for which our convergence analysis is valid), but we also use the conjugate gradient method (without theoretical support). Finally, we compare with the well-known “fast iterative shrinkage-thresholding algorithm” (FISTA). The numerical results demonstrate that our new step size rule improves previous step size rules for the superiorization methodology and is competitive with, and in several instances behaves better than, the other methods.

A joint convex penalty for inverse covariance matrix estimation

Conjugate gradient algorithms come in a wide range of flavors. Conjugate gradient techniques primarily concentrate on the coefficient conjugate. We introduce a novel conjugate gradient approach that computes the parameter by using Newton updates. In addition, we have demonstrated that our conjugate gradient algorithms are globally convergent and descent property. For the specified test issues in [1] the performance profiles revealed that the novel conjugate gradient approach is effective and efficient.

In this paper, we introduce an augmented Lagrangian for nonlinear semidefinite programs. Some basic properties of the augmented Lagrangian such as differentiabilty, monotonicity and convexity, are discussed. Necessary and sufficient conditions for a strong duality property and an exact penalty representation in the framework of augmented Lagrangian are derived. Under certain conditions, it is shown that any limit point of a sequence of stationary points of augmented Lagrangian problems is a Karuh, Kuhn-Tucker (for short, KKT) point of the original semidefinite program.

Energy-efficient power allocation in cell-free massive MIMO with zero-forcing: First order methods

A key component of conjugate gradient (CG) algorithms is the directions used in the CG approach. Because of their low memory requirements, these directed methods have proven effective in various applications, especially image processing. For this study, we have developed a new conjugate gradient coefficient based on the well-known quadratic model, that will accelerate the convergence and increase the accuracy of the CG approach. Because the new method uses the second order curvature and provides improved direction. So, the global convergence and critical descent characteristics of the final algorithm provide reliable performance in a range of scenarios. The results of extensive numerical testing on picture restoration show that the new formulae perform noticeably better than the current techniques. In particular, the performance of the suggested conjugate gradient (CG) scheme has been better than that of the conventional Fletcher-Reeves (FR) conjugate gradient approach. This development is a key tool in the field of computational imaging and optimization as it not only increases computing efficiency but also boosts picture restoration, image quality and its' accuracy.

Compressed sensing (CS) in magnetic resonance imaging (MRI) enables the reconstruction of MR images from highly undersampled k-spaces, and thus substantial reduction of data acquisition time. In this context, edge-preserving and sparsity-promoting regularizers are used to exploit the prior knowledge that MR images are sparse or compressible in a given transform domain and thus to regulate the solution space. In this study, we introduce a new regularization scheme by iterative linearization of the non-convex clipped absolute deviation (SCAD) function in an augmented Lagrangian framework. The performance of the proposed regularization, which turned out to be an iteratively weighted total variation (TV) regularization, was evaluated using 2D phantom simulations and 3D retrospective undersampling of clinical MRI data by different sampling trajectories. It was demonstrated that the proposed regularization technique substantially outperforms conventional TV regularization, especially at reduced sampling rates.

In this paper, we consider a composite optimization problem with linear coupling constraints in a multi-agent network. In this problem, the agents cooperatively optimize a strongly convex cost function which is the linear sum of individual cost functions composed of smooth and possibly non-smooth components. To solve this problem, we propose an asynchronous penalized proximal gradient (Asyn-PPG) algorithm, a variant of classical proximal gradient method, with the presence of the asynchronous updates of the agents and uniform communication delays in the network. Specifically, we consider a slot-based asynchronous network (SAN), where the whole time domain is split into sequential time slots and each agent is permitted to execute multiple updates during a slot by accessing the historical state information of the agents. By the Asyn-PPG algorithm, an explicit convergence rate can be guaranteed based on deterministic analysis. The feasibility of the proposed algorithm is verified by solving a consensus-based distributed regression problem and a social welfare optimization problem in the electricity market.

SPIRiT (iterative self-consistent parallel imaging reconstruction), and its sparsity-regularized variant L1-SPIRiT, are compatible with both Cartesian and non-Cartesian magnetic resonance imaging sampling trajectories. However, the non-Cartesian framework is more expensive computationally, involving a nonuniform Fourier transform with a nontrivial Gram matrix. We propose a novel implementation of the regularized reconstruction problem using variable splitting, alternating minimization of the augmented Lagrangian, and careful preconditioning. Our new method based on the alternating direction method of multipliers converges much faster than existing methods because of the preconditioners' heightened effectiveness. We demonstrate such rapid convergence substantially improves image quality for a fixed computation time. Our framework is a step forward towards rapid non-Cartesian L1-SPIRiT reconstructions.

In this work, we propose a deep learning approach for parallel magnetic resonance imaging (MRI) reconstruction, termed a variable splitting network (VS-Net), for an efficient, high-quality reconstruction of undersampled multi-coil MR data. We formulate the generalized parallel compressed sensing reconstruction as an energy minimization problem, for which a variable splitting optimization method is derived. Based on this formulation we propose a novel, end-to-end trainable deep neural network architecture by unrolling the resulting iterative process of such variable splitting scheme. VS-Net is evaluated on complex valued multi-coil knee images for 4-fold and 6-fold acceleration factors. We show that VS-Net outperforms state-of-the-art deep learning reconstruction algorithms, in terms of reconstruction accuracy and perceptual quality. Our code is publicly available at https://github.com/j-duan/VS-Net.

Algorithms to solve constrained optimization problems are derived. These schemes combine an unconstrained minimization scheme like the conjugate gradient method, an augmented Lagrangian, and multiplier updates to obtain global quadratic convergence. Since an augmented Lagrangian can be ill conditioned, a preconditioning strategy is developed to eliminate the instabilities associated with the penalty term. A criterion for deciding when to increase the penalty is presented.

Purpose– Generally, iterative methods which have some instability solutions in complex structural and non-linear mechanical problems are used to compute reliability index. The purpose of this paper is to establish a non-linear conjugate gradient (NCG) optimization algorithm to overcome instability solution of the Hasofer-Lind and Rackwitz-Fiessler (HL-RF) method in first-order reliability analysis. The NCG algorithms such as the Conjugate-Descent (CD) and the Liu-Storey (LS) are used for determining the safety index. An algorithm is found based on the new line search in the reliability analysis.Design/methodology/approach– In the proposed line search for calculating the safety index, search direction is computed by using the conjugate gradient approach and the HL-RF method based on the new and pervious gradient vector of the reliability function. A simple step size is presented for the line search in the proposed algorithm, which is formulated by the Wolfe conditions based on the new and previous safety index results in the reliability analysis.Findings– From the current work, it is concluded that the proposed NCG algorithm has more efficient, robust and appropriate convergence in comparison with the HL-RF method. The proposed methods can eliminate numerical instabilities of the HL-RF iterative algorithm in highly non-linear performance function and complicated structural limit state function. The NGC optimization is applicable to reliability analysis and it is correctly converged on the reliability index. In the NCG method, the CD algorithm is slightly more efficient than the LS algorithm.Originality/value– This paper usefully shows how the HL-RF algorithm and the NCG scheme are formulated in first-order reliability analysis. The proposed algorithm is validated from six numerical and structural examples taken from the literature. The HL-RF method is not converged on several non-linear mathematic and complex structural examples, while the two proposed conjugate gradient methods are appropriately converged for all examples. The CD algorithm is converged about twice faster than the LS algorithm in most of the problems. Therefore, application of the NCG method is possible in reliability analysis.