Query：
学者姓名：夏又生
Refining：
Year
Type
Indexed by
Source
Complex
Co-
Language
Clean All
Abstract ：
提出一种基于矩阵型神经动力学优化的非负矩阵分解算法.将矩阵非负分解优化问题首先转换为两个矩阵变量凸优化子问题,针对其子问题分别提出矩阵型惯性投影神经网络;然后,采用交替迭代方案寻找矩阵非负分解优化问题的解.理论分析证明了矩阵型惯性投影神经网络能收敛于矩阵变量凸优化子问题的最优解,并且基于矩阵型神经网络的交替迭代算法可以收敛到矩阵非负分解优化问题的偏最优解.最后,所提出的基于矩阵型神经网络的交替迭代算法被有效地应用于人脸识别.
Keyword ：
人脸识别 人脸识别 惯性投影神经网络 惯性投影神经网络 矩阵动力学优化 矩阵动力学优化 非负矩阵分解 非负矩阵分解
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 李小玲 , 夏又生 . 基于矩阵型惯性投影神经网络的非负矩阵分解算法 [J]. | 福州大学学报（自然科学版） , 2023 , 51 (1) : 1-8 . |
| MLA | 李小玲 等. "基于矩阵型惯性投影神经网络的非负矩阵分解算法" . | 福州大学学报（自然科学版） 51 . 1 (2023) : 1-8 . |
| APA | 李小玲 , 夏又生 . 基于矩阵型惯性投影神经网络的非负矩阵分解算法 . | 福州大学学报（自然科学版） , 2023 , 51 (1) , 1-8 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
提出一种基于L0范数正则化的自然图像去反光算法.首先,根据自然反光图像的两个特征构建基于L0范数的正则优化模型,保证漫反射图像系数矩阵的稀疏性、低秩性和反光区域漫反射分量的有效恢复.其次,利用增广拉格朗日技术,导出求解L0范数正则优化模型的算法.最后,通过与相关的图像去反光算法对比,证实本图像去反光算法在均方误差和结构相似度上均优于其他去反光算法,使其生成图像在保留更多纹理细节信息的同时,可以有效去除图像反光.
Keyword ：
L0范数正则化 L0范数正则化 图像反光去除 图像反光去除 矩阵变量优化 矩阵变量优化
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 丁凤 , 夏又生 . 基于L0矩阵范数正则化的自然图像去反光算法 [J]. | 福州大学学报（自然科学版） , 2022 , 50 (6) : 729-736 . |
| MLA | 丁凤 等. "基于L0矩阵范数正则化的自然图像去反光算法" . | 福州大学学报（自然科学版） 50 . 6 (2022) : 729-736 . |
| APA | 丁凤 , 夏又生 . 基于L0矩阵范数正则化的自然图像去反光算法 . | 福州大学学报（自然科学版） , 2022 , 50 (6) , 729-736 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
In recent years, matrix-valued optimization algorithms have been studied to enhance the computational performance of vector-valued optimization algorithms. This paper presents two matrix-type projection neural networks, continuous-time and discrete-time ones, for solving matrix-valued optimization problems. The proposed continuous-time neural network may be viewed as a significant extension to the vector-type double projection neural network. More importantly, the proposed discrete-time projection neural network is suitable for parallel implementation in terms of matrix state spaces. Under pseudo-monotonicity and Lipschitz continuous conditions, the proposed two matrix-type projection neural networks are guaranteed to be globally convergent to the optimal solution. Finally, the proposed matrix-type projection neural network is effectively applied to image restoration. Computed examples show that the two proposed matrix-type projection neural networks are much superior to the vector-type projection neural networks in terms of computation speed.
Keyword ：
Fast computation Fast computation Global convergence Global convergence Image restoration Image restoration Matrix-type neural network Matrix-type neural network Matrix-valued optimization Matrix-valued optimization
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Huang, Lingmei , Xia, Youshen , Huang, Liqing et al. Two Matrix-Type Projection Neural Networks for Matrix-Valued Optimization with Application to Image Restoration [J]. | NEURAL PROCESSING LETTERS , 2021 , 53 (3) : 1685-1707 . |
| MLA | Huang, Lingmei et al. "Two Matrix-Type Projection Neural Networks for Matrix-Valued Optimization with Application to Image Restoration" . | NEURAL PROCESSING LETTERS 53 . 3 (2021) : 1685-1707 . |
| APA | Huang, Lingmei , Xia, Youshen , Huang, Liqing , Zhang, Songchuan . Two Matrix-Type Projection Neural Networks for Matrix-Valued Optimization with Application to Image Restoration . | NEURAL PROCESSING LETTERS , 2021 , 53 (3) , 1685-1707 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Super resolution image reconstruction under unknown Gaussian blur has been a challenging topic. Advanced optimization-based works for blind image super-resolution (SR) were reported to be effective, but there exist both large data space storage and time consuming due to vector-variable optimization. This paper proposes a matrix-variable optimization method for fast blind image SR. We first present an accurate blur kernel estimation-based matrix decomposition method. Then we propose minimizing a matrix-variable optimization problem with sparse representation and TV regularization terms. The proposed method can exactly estimate the unknown blur kernel and blur matrix. Compared with vector-variable optimization based methods for blind image SR, the proposed method can greatly reduce their data space storage and computation time. Compared with deep learning methods, the proposed method can directly deal with multiframe SR problem without training and learning task. Experimental results show that the proposed algorithm is superior to conventional optimization-based method in terms of solution quality and computation time. Moreover, the proposed method can obtain higher reconstruction quality than the deep learning methods, specially in the case of large blur kernels.
Keyword ：
fast computation fast computation Image reconstruction Image reconstruction Image resolution Image resolution Image super-resolution Image super-resolution Kernel Kernel Matrix decomposition Matrix decomposition matrix-variable optimization matrix-variable optimization Optimization methods Optimization methods reconstruction quality reconstruction quality Sparse matrices Sparse matrices unknown blur unknown blur
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Huang, Liqing , Xia, Youshen . Fast Blind Image Super Resolution Using Matrix-Variable Optimization [J]. | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY , 2021 , 31 (3) : 945-955 . |
| MLA | Huang, Liqing et al. "Fast Blind Image Super Resolution Using Matrix-Variable Optimization" . | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 31 . 3 (2021) : 945-955 . |
| APA | Huang, Liqing , Xia, Youshen . Fast Blind Image Super Resolution Using Matrix-Variable Optimization . | IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY , 2021 , 31 (3) , 945-955 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Convolutional neural networks (CNN) have shown its excellent performance in computer vision fields. Recently, they are successfully applied to image restoration. This paper proposes a joint blur kernel estimation and CNN method for blind image restoration. The blur kernel estimation is based on both blur support parameter estimation and blur type identification. An automatic feature line detection algorithm is presented for blur support parameter estimation and a dictionary learning algorithm is presented for the blur type identification. Once the blur kernel estimate is obtained, we use an effective CNN for iterative non-blind deconvolution, which is able to automatically learn image priors. Compared with current blind image restoration methods, the proposed joint method can obtain restored images under three types of unknown blur kernels. The experimental result shows that the proposed blur kernel estimation algorithm can provide high accuracy results. Furthermore, the proposed joint blur kernel estimation and CNN algorithm is superior to conventional blind image restoration algorithms in terms of restoration quality and computation time. (C) 2019 Elsevier B.V. All rights reserved.
Keyword ：
Blind image restoration Blind image restoration Blur kernel Blur kernel Blur support parameter estimation Blur support parameter estimation Blur type identification Blur type identification CNN CNN
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Huang, Liqing , Xia, Youshen . Joint blur kernel estimation and CNN for blind image restoration [J]. | NEUROCOMPUTING , 2020 , 396 : 324-345 . |
| MLA | Huang, Liqing et al. "Joint blur kernel estimation and CNN for blind image restoration" . | NEUROCOMPUTING 396 (2020) : 324-345 . |
| APA | Huang, Liqing , Xia, Youshen . Joint blur kernel estimation and CNN for blind image restoration . | NEUROCOMPUTING , 2020 , 396 , 324-345 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Recent reports show that projection neural networks with a low-dimensional state space can enhance computation speed obviously. This paper proposes two projection neural networks with reduced model dimension and complexity (RDPNNs) for solving nonlinear programming (NP) problems. Compared with existing projection neural networks for solving NP, the proposed two RDPNNs have a low-dimensional state space and low model complexity. Under the condition that the Hessian matrix of the associated Lagrangian function is positive semi-definite and positive definite at each Karush-Kuhn-Tucker point, the proposed two RDPNNs are proven to be globally stable in the sense of Lyapunov and converge globally to a point satisfying the reduced optimality condition of NP. Therefore, the proposed two RDPNNs are theoretically guaranteed to solve convex NP problems and a class of nonconvex NP problems. Computed results show that the proposed two RDPNNs have a faster computation speed than the existing projection neural networks for solving NP problems.
Keyword ：
Computational complexity Computational complexity Computational modeling Computational modeling Convex programming Convex programming fast computation fast computation global stability global stability low-dimensional state space low-dimensional state space Manganese Manganese Neural networks Neural networks nonconvex programming nonconvex programming Optimization Optimization Programming Programming
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Xia, Youshen , Wang, Jun , Guo, Wenzhong . Two Projection Neural Networks With Reduced Model Complexity for Nonlinear Programming [J]. | IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS , 2020 , 31 (6) : 2020-2029 . |
| MLA | Xia, Youshen et al. "Two Projection Neural Networks With Reduced Model Complexity for Nonlinear Programming" . | IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 31 . 6 (2020) : 2020-2029 . |
| APA | Xia, Youshen , Wang, Jun , Guo, Wenzhong . Two Projection Neural Networks With Reduced Model Complexity for Nonlinear Programming . | IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS , 2020 , 31 (6) , 2020-2029 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Recently, a matrix-type neural dynamical method for matrix-variable nonlinear optimization with box constraints was presented. This paper proposes two matrix-type neural dynamical optimization methods for matrix-variable nonlinear programming with linear constraints. Each matrix-type neural dynamical method consists of continuous-time and discrete-time models. The two continuous-time models significantly generalize two existing vector-type projection neural networks, while the two discrete-time state models have low complexity and can be implemented parallelly by matrix operation. Under proper conditions, the proposed two matrix-type neural dynamical methods are guaranteed to converge globally to the optimal solution. Finally, computed examples show that the proposed matrix-type neural dynamical methods for matrix-variable nonlinear programming with linear constraints are superior to current matrix-type neural dynamical methods in fast computation. © 2020 IEEE.
Keyword ：
Continuous time systems Continuous time systems Intelligent computing Intelligent computing Matrix algebra Matrix algebra Nonlinear programming Nonlinear programming
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Ye, Tiantian , Xia, Youshen . Matrix-type neural dynamical methods for matrix-variable nonlinear programming with linear constraints [C] . 2020 : 23-29 . |
| MLA | Ye, Tiantian et al. "Matrix-type neural dynamical methods for matrix-variable nonlinear programming with linear constraints" . (2020) : 23-29 . |
| APA | Ye, Tiantian , Xia, Youshen . Matrix-type neural dynamical methods for matrix-variable nonlinear programming with linear constraints . (2020) : 23-29 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Stereo matching aims to obtain depth information of scene from captured images, which becomes an active research topic in the field of computer vision. Most stereo matching cost algorithms are based on a common assumption, which is the intensity or color value of corresponding pixels are same. However, in real-world applications, the colors of the objects observed in the recorded image data are affected by radiometric variations. In this paper, using a novel similarity measure we propose a robust stereo matching method, which has robust performance to noise, illumination condition changes, and exposure changes between left and right images. The proposed stereo matching cost combines improved zero mean normalized cross-correlation (ZNCC) model and the absolute difference of local binary pattern (LBP) of windows to get both the color and texture similarity of windows to be matched. Based on Middleburry data set, we verify the effectiveness of the proposed algorithm. Computed results show that the proposed algorithm is more robust to illumination changes and noise than related stereo matching algorithms. © 2020 Association for Computing Machinery.
Keyword ：
Color Color Color matching Color matching Radiometry Radiometry Stereo image processing Stereo image processing Stereo vision Stereo vision Textures Textures
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Chen, Jieqiong , Xia, Yousheng . Robust stereo matching using improved ZNCC combined SAD-LBP [C] . 2020 : 141-146 . |
| MLA | Chen, Jieqiong et al. "Robust stereo matching using improved ZNCC combined SAD-LBP" . (2020) : 141-146 . |
| APA | Chen, Jieqiong , Xia, Yousheng . Robust stereo matching using improved ZNCC combined SAD-LBP . (2020) : 141-146 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Variational Retinex model-based methods for low-light image enhancement have been popularly studied in recent years. In this paper, we present an enhanced variational Retinex method for low-light natural image enhancement, based on the initial smoother illumination component with a structure extraction technique. The Bergman splitting algorithm is then introduced to estimate the illuminance component and reflectance component. The de-block processing and illuminance component correction are used for the enhanced reflectance as the ultimate enhanced image. Moreover, the estimated smoother illumination component can make enhanced images preserve edge details. Experimental results with a comparison demonstrate the present variational Retinex method can effectively enhance image quality and maintain image color. © 2020, Springer Nature Switzerland AG.
Keyword ：
Computer vision Computer vision Extraction Extraction Image enhancement Image enhancement Lighting Lighting Reflection Reflection
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Du, Xiaoyu , Xia, Youshen . Natural Images Enhancement Using Structure Extraction and Retinex [C] . 2020 : 408-420 . |
| MLA | Du, Xiaoyu et al. "Natural Images Enhancement Using Structure Extraction and Retinex" . (2020) : 408-420 . |
| APA | Du, Xiaoyu , Xia, Youshen . Natural Images Enhancement Using Structure Extraction and Retinex . (2020) : 408-420 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Much research has been devoted to the problem of low-dose computerized tomography (CT) image reconstruction with a scan protocol by both lowering the X-ray tube current (low-mAs) and reducing the total number of projection views(sparse view). However, the CT transmission data may be severely corrupted by X-ray quanta noise and system electronic noise. Recently, a non local means (NLM) regularization method for low-dose CT reconstruction was proposed. Although this method is effective in suppressing both Poisson Gaussian noise and artifacts, it has two disadvantages: the heavy computational burden and blurred edge information in CT reconstruction. This paper proposes an adaptive patch-wise regularization method for low-dose CT reconstruction from available CT transmission data in Poisson Gaussian noise. The proposed cost function includes a penalized weighted least-square term and two patch-wise regularization terms which are combined with a novel adaptive regularization parameter. The two regularization terms take advantage of image redundant information across different scales. By exploiting both global and local structure information, the reconstruction accuracy is enhanced. In addition, we select the adaptive regularization parameter based on the texture of the image patch so that the edge information is further maintained. Furthermore, our algorithm updates a patch rather than a pixel, the computation burden is greatly reduced. Finally, experiment results show that the proposed adaptive patch-wise regularization method for low-dose CT reconstruction is superior to several conventional regularization-based approaches in terms of computation efficiency and resolution quality.
Keyword ：
Adaptive regularization method Adaptive regularization method Low-dose Low-dose Multi-scale Multi-scale Poisson Gaussian noise Poisson Gaussian noise
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Zhang, Shu , Xia, Youshen , Zou, Changzhong . An adaptive regularization method for low-dose CT reconstruction from CT transmission data in Poisson-Gaussian noise [J]. | OPTIK , 2019 , 188 : 172-186 . |
| MLA | Zhang, Shu et al. "An adaptive regularization method for low-dose CT reconstruction from CT transmission data in Poisson-Gaussian noise" . | OPTIK 188 (2019) : 172-186 . |
| APA | Zhang, Shu , Xia, Youshen , Zou, Changzhong . An adaptive regularization method for low-dose CT reconstruction from CT transmission data in Poisson-Gaussian noise . | OPTIK , 2019 , 188 , 172-186 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Export
| Results： |
Selected to |
| Format： |
