We propose a new algorithm for the blind source separation of acoustic sources. This algorithm is an alternative to the popular auxiliary function based independent vector analysis using iterative projection (AuxIVA-I...
详细信息
ISBN:
(纸本)9781509066315
We propose a new algorithm for the blind source separation of acoustic sources. This algorithm is an alternative to the popular auxiliary function based independent vector analysis using iterative projection (AuxIVA-IP). It optimizes the same cost function, but instead of alternate updates of the rows of the demixing matrix, we propose a sequence of rank-1 updates. Remarkably, and unlike the previous method, the resulting updates do not require matrix inversion. Moreover, their computational complexity is quadratic in the number of microphones, rather than cubic in AuxIVA-IP. In addition, we show that the new method can be derived as alternate updates of the steering vectors of sources. Accordingly, we name the method iterative source steering (AuxIVA-ISS). Finally, we confirm in simulated experiments that the proposed algorithm separates sources just as well as AuxIVA-IP, at a lower computational cost.
This paper is concerned with long time numerical behaviors of nonlinear fractional pantograph equations. The L1 method with the linear interpolation procedure is applied to solve these nonlinear problems. It is proved...
详细信息
This paper is concerned with long time numerical behaviors of nonlinear fractional pantograph equations. The L1 method with the linear interpolation procedure is applied to solve these nonlinear problems. It is proved that the proposed numerical scheme can inherit the long time behavior of the underlying problems without any stepsize restrictions. After that, the fast evaluation is presented to speed up the calculation of the Caputo fractional derivative. Numerical examples are shown to confirm the theoretical results. Besides, several counter-examples are also given to show that not all the numerical methods can inherit the long time behavior of the underlying problems. (C) 2019 International Association for Mathematics and Computers in Simulation (IMACS). Published by Elsevier B.V. All rights reserved.
In this paper, a novel L-2-norm estimation method for blind data fusion under noisy environments is proposed and a fast learning algorithm is developed to implement the proposed estimation method. The proposed learnin...
详细信息
In this paper, a novel L-2-norm estimation method for blind data fusion under noisy environments is proposed and a fast learning algorithm is developed to implement the proposed estimation method. The proposed learning algorithm is proved to be globally exponentially convergent to an optimal fusion weight vector. In addition, the proposed learning algorithm has lower computation complexity than the existing cooperative learning algorithm based a L-1-norm estimation method. Compared with other estimation methods, the proposed estimation method can be effectively used in the blind image fusion. Application examples of image fusion show that the proposed learning algorithm is able to fast obtain more accurate solutions than several conventional algorithms.
A fast element-free Galerkin (EFG) method is proposed for the numerical analysis of the fractional diffusion-wave equation. In this method, a fast time discrete scheme is first derived by applying the L1 and the fast ...
详细信息
A fast element-free Galerkin (EFG) method is proposed for the numerical analysis of the fractional diffusion-wave equation. In this method, a fast time discrete scheme is first derived by applying the L1 and the fast H2N2 approximations to discretize the time Caputo fractional derivative, and then the Nitsche's technique and the stabilized moving least squares approximation are adopted to establish linear algebraic systems. Based on the reproducing kernel gradient smoothing integration, an efficient numerical integration procedure is also presented to further accelerate the computations of the method. Theoretical error analysis of the fast EFG method is provided. Numerical results verify the efficiency of the method. (C) 2021 Elsevier Ltd. All rights reserved.
In this paper a Multistep Collocation (MC) method for the second kind Fredholm integral equations (FIEs) is proposed and analyzed. The multistep collocation method is applied to FIE with smooth kernels under uniform m...
详细信息
In this paper a Multistep Collocation (MC) method for the second kind Fredholm integral equations (FIEs) is proposed and analyzed. The multistep collocation method is applied to FIE with smooth kernels under uniform mesh and weakly singular kernels vertical bar s - t vertical bar(-alpha) (0 < alpha < 1) using a graded mesh then the same convergence rate as collocation method but with a lower degree of freedom is obtained. Moreover, in order to avoid the round-off errors caused by graded mesh, a Hybrid Multistep Collocation (HMC) method by combining multistep collocation and hybrid collocation method is proposed. The HMC method converges faster with lower degrees of freedom and more efficiently captures the weakly singular properties by nonpolynomial interpolation at the first subinterval. The L-infinity-norm convergence results are analysed and proved. Numerical examples are presented to demonstrate the efficiency of the proposed methods. (C) 2021 Elsevier Inc. All rights reserved.
A novel solve-training framework is proposed to train neural network in representing low dimensional solution maps of physical models. Solve-training framework uses the neural network as the ansatz of the solution map...
详细信息
A novel solve-training framework is proposed to train neural network in representing low dimensional solution maps of physical models. Solve-training framework uses the neural network as the ansatz of the solution map and trains the network variationally via loss functions from the underlying physical models. Solve-training framework avoids expensive data preparation in the traditional supervised training procedure, which prepares labels for input data, and still achieves effective representation of the solution map adapted to the input data distribution. The efficiency of solve-training framework is demonstrated through obtaining solution maps for linear and nonlinear elliptic equations, and maps from potentials to ground states of linear and nonlinear Schrodinger equations. (C) 2020 Elsevier Inc. All rights reserved.
In recent years, a number of fast algorithms for computing the determinant of a Toeplitz matrix were developed. The fastest algorithm we know so far is of order k(2) log n + k(3), where n is the number of rows of the ...
详细信息
In recent years, a number of fast algorithms for computing the determinant of a Toeplitz matrix were developed. The fastest algorithm we know so far is of order k(2) log n + k(3), where n is the number of rows of the Toeplitz matrix and k is the bandwidth size. This is possible because such a determinant can be expressed as the determinant of certain parts of the n-th power of a related k x k companion matrix. In this paper, we give a new elementary proof of this fact, and provide various examples. We give symbolic formulas for the determinants of Toeplitz matrices in terms of the eigenvalues of the corresponding companion matrices when k is small. (C) 2013 Elsevier B.V. All rights reserved.
Static test plays a key role in accuracy evaluation in range equipment test. However, the ground station equipment in static test may inevitable encounter intermittent fault, which leads to inaccurate navigation resul...
详细信息
ISBN:
(纸本)9781728176871
Static test plays a key role in accuracy evaluation in range equipment test. However, the ground station equipment in static test may inevitable encounter intermittent fault, which leads to inaccurate navigation result. In this paper, two improved intermittent fault algorithms are proposed by using sorting, iteration, and recursion techniques based on multivariate statistical analysis. These algorithms not only reduce significantly the computational complexity of fault detection based on the traditional detection statistics, but also improve the navigation accuracy. It will have a wide application prospect in the future flight test and final equipment test.
The fault diagnosis of bevel gearbox is of great significance. At present, the commonly used methods are based on pattern recognition, such as support vector machine, convex hull classifier and hyperdisk classifier. H...
详细信息
The fault diagnosis of bevel gearbox is of great significance. At present, the commonly used methods are based on pattern recognition, such as support vector machine, convex hull classifier and hyperdisk classifier. However, the number of elements in the kernel matrix of these kernel function-based classification methods increases squarely with the data size, resulting in intolerable training time. Based on this, a sparse random projection-based hyperdisk classifier model is proposed. The proposed method has the following novelties: First, based on sparse random projection and the geometrical characteristics of the hyperdisk model, a method is designed to efficiently screen out the core samples, and these samples are given different weights in this process. Second, the proposed method introduces slack variables and the dynamic penalty parameter to obtain a hyperdisk model with more reasonable boundary. Last, a strategy is developed to minimize the adverse effects of imbalanced training data. The effectiveness and applicability of the proposed method are verified on bevel gearbox fault data. The experimental results show that compared with other classifiers, the proposed method can greatly reduce the training time while guaranteeing a high classification accuracy. What's more, it has better performance and efficiency in fault diagnosis with imbalanced training data.
Region duplication is one of the most common methods of video forgery. Existing forgery detection algorithms generally suffer from inefficiency and are not effective for the forged regions with mirroring. To address t...
详细信息
Region duplication is one of the most common methods of video forgery. Existing forgery detection algorithms generally suffer from inefficiency and are not effective for the forged regions with mirroring. To address these problems, we present a fast forgery detection algorithm based on Exponential Fourier moments (EFMs) for detecting region duplication in videos. The algorithm first extracts EFMs features from each block in the current frame, and performs a fast match to find potential matching pairs. Then, a postverification scheme is designed to eliminate falsely matched pairs and locate the altered regions in the current frame. Finally, an adaptive parameter-based fast compression tracking algorithm is used to track the tampered regions in the subsequent frames. The experimental results show that our proposed algorithm has higher detection accuracy and computational efficiency than those of previous algorithms.
暂无评论