检索结果-内蒙古大学图书馆

Learned Block iterative shrinkage thresholding algorithm for Photothermal Super Resolution Imaging

SENSORS 2022年第15期22卷 5533-5533页

作者： Hauffen, Jan Christian Kaestner, Linh Ahmadi, Samim Jung, Peter Caire, Giuseppe Ziegler, Mathias Berlin Inst Technol Commun & Informat Theory D-10623 Berlin Germany Berlin Inst Technol Fac Elect Engn & Comp Sci Ind Grade Networks & Clouds D-10623 Berlin Germany Bundesanstalt Mat Forsch & Prufung Dept Nondestruct Testing D-12489 Berlin Germany

Block-sparse regularization is already well known in active thermal imaging and is used for multiple-measurement-based inverse problems. The main bottleneck of this method is the choice of regularization parameters which differs for each experiment. We show the benefits of using a learned block iterative shrinkage thresholding algorithm (LBISTA) that is able to learn the choice of regularization parameters, without the need to manually select them. In addition, LBISTA enables the determination of a suitable weight matrix to solve the underlying inverse problem. Therefore, in this paper we present LBISTA and compare it with state-of-the-art block iterative shrinkage thresholding using synthetically generated and experimental test data from active thermography for defect reconstruction. Our results show that the use of the learned block-sparse optimization approach provides smaller normalized mean square errors for a small fixed number of iterations. Thus, this allows us to improve the convergence speed and only needs a few iterations to generate accurate defect reconstruction in photothermal super-resolution imaging.

关键词： active thermal imaging block-sparsity deep unfolding defect reconstruction iterative shrinkage thresholding algorithm laser thermography learned optimization neural network regularization

来源：评论

学校读者我要写书评

暂无评论

Structured LISTA for Multidimensional Harmonic Retrieval

引用

IEEE TRANSACTIONS ON SIGNAL PROCESSING 2021年 69卷 3459-3472页

作者： Fu, Rong Liu, Yimin Huang, Tianyao Eldar, Yonina C. Tsinghua Univ Dept Elect Engn Beijing 100084 Peoples R China Weizmann Inst Sci Fac Math & CS IL-7610001 Rehovot Israel

Learned iterative shrinkage thresholding algorithm (LISTA), which adopts deep learning techniques to optimize algorithm parameters from labeled training data, can be successfully applied to small-scale multidimensional harmonic retrieval (MHR) problems. However, LISTA becomes computationally demanding for large-scale MHR because the matrix size of the learned mutual inhibition matrix exhibits quadratic growth with the signal length. These large matrices consume costly memory/computation resources and require a huge amount of labeled data for training. For MHR problems, the mutual inhibition matrix naturally has a Toeplitz structure, implying the degrees of freedom of the matrix can be reduced from quadratic order to linear order. We thereby propose a structured LISTA-Toeplitz network, which imposes Toeplitz structure on the mutual inhibition matrices and applies linear convolution instead of matrix-vector multiplications in traditional LISTA. Both simulation and field tests for air target detection with radar are carried out to validate the performance of the proposed network. For small-scale MHR problems, LISTA-Toeplitz exhibits close or even better recovery accuracy than traditional LISTA, while the former significantly reduces the network complexity and requires much less training data. For large-scale MHR problems, where LISTA is difficult to implement due to the huge size of the matrices, our proposed LISTA-Toeplitz still enjoys good recovery performance.

关键词： Harmonic analysis Convolution Dictionaries Sparse matrices Training Signal processing algorithms Estimation Compressed sensing multidimensional harmonic retrieval iterative shrinkage thresholding algorithm learned ISTA Toeplitz structure

来源：评论

学校读者我要写书评

暂无评论

Improved SAR imaging algorithm with azimuth periodically missing data

引用

IET RADAR SONAR AND NAVIGATION 2020年第3期14卷 399-406页

作者： Yang, Weixing Bi, Hui Zhu, Daiyin Nanjing Univ Aeronaut & Astronaut Key Lab Radar Imaging & Microwave Photon Coll Elect & Informat Engn 29 Jiangjun Ave Nanjing Peoples R China Nanjing Univ Aeronaut & Astronaut Minist Educ Coll Elect & Informat Engn 29 Jiangjun Ave Nanjing Peoples R China

Aiming at the periodically missing data of synthetic aperture radar (SAR) in azimuth, a reconstruction and imaging algorithm based on ${\bi L}_p$Lp-alternating direction method with compressed sensing theory is proposed. The algorithm can effectively suppress ghosting and aliasing caused by azimuth missing data, and improve the imaging quality. To reduce memory consumption and lower computational complexity, approximate observation model based on SAR raw data simulator is utilised to rapid reconstruction imaging. Compared to the traditional iterative shrinkage thresholding algorithm, the proposed algorithm has better reconstruction image quality. Simulation and raw SAR echo data processing demonstrate the effectiveness of the proposed method in solving the problem of imaging with periodically missing data in SAR azimuth.

关键词： image reconstruction computational complexity synthetic aperture radar iterative methods radar imaging compressed sensing data processing SAR azimuth SAR imaging algorithm synthetic aperture radar compressed sensing theory ghosting aliasing azimuth missing data imaging quality memory consumption approximate observation model SAR raw data rapid reconstruction imaging iterative shrinkage thresholding algorithm image reconstruction Lp-alternating direction method

来源：评论

学校读者我要写书评

暂无评论

Separation of simultaneous source data via an accelerated joint iterative method

引用

JOURNAL OF APPLIED GEOPHYSICS 2022年 203卷

作者： Zhang, Hua Liang, Shuang Peng, Qing Pang, Yang Jiang, Ying Song, Wenlong East China Univ Technol State Key Lab Nucl Resources & Environm Nanchang 330013 Peoples R China

At present, there are many deblending algorithms to deal with the separation of the blended data. However, the convergence speed for most algorithms is relatively slow, which is difficult to implement in industrial production. Thus, we propose an accelerated joint iterative method (AJIM) to separate the blended data based on the iterative shrinkage thresholding algorithm (ISTA) and the acceleration linearized Bregman method (ALBM) in the curvelet transform domain. In conventional ISTA, the updated deblended result of each iteration is obtained by thresholding curvelet coefficients and the gradient term. It is stable but always needs a lot of iterations to obtain a good performance. ALBM is proposed based on LBM via an acceleration factor at each iteration, and the acceleration factor can increase the proportion of unthresholded curvelet coefficient, which further accelerates the convergence. This method converges more faster than LBM and ISTA at the beginning, but it can cause artifacts at the late iterations because some unthresholded coefficients contain some wrong information. Based on advantages of ISTA and ALBM, the weight parameters with linear and exponential functions are proposed to control the contribution of ISTA and ALBM in the proposed AJIM. Therefore, AJIM can greatly speed up the deblending efficiency at the beginning and improve the deblending accuracy in the late stage. At last, AJIM can obtain the same deblending effect as ISTA, but only a few iterations are used. Synthetic and field data tests are used to demonstrate that AJIM-based deblending can effectively and quickly separate the blended data.

关键词： Deblending Linearized Bregman method (LBM) iterative shrinkage thresholding algorithm (ISTA) Simultaneous

来源：评论

学校读者我要写书评

暂无评论

Adaptive vectorial total variation models for multi-channel synthetic aperture radar images despeckling with fast algorithms

引用

IET IMAGE PROCESSING 2013年第9期7卷 795-804页

作者： Liu, Huiyan Yan, Fengxia Zhu, Jubo Fang, Faming Natl Univ Def Technol Dept Math & Syst Sci Changsha 410073 Hunan Peoples R China E China Normal Univ Dept Comp Sci Shanghai 200241 Peoples R China

This study proposes two adaptive vectorial total variation models for multi-channel synthetic aperture radar (SAR) images despeckling with the help of prior knowledge of the image amplitude. Besides despeckling the multi-channel SAR images efficiently, the proposed new models have advantages over other total variation methods in many aspects, such as preserving the radar reflectivity, the targets and edges contrast. The Bermudez-Moreno algorithm and the accelerated fast iterative shrinkage thresholding algorithm are employed to implement the new two models, respectively. Experimental results on multi-polarimetric, multi-temporal RADARSAT-2 images show that the visual quality and evaluation indexes of the proposed models and the corresponding algorithms outperform the other methods with edge preservation.

关键词： image segmentation iterative methods radar imaging synthetic aperture radar vectors adaptive vectorial total variation models multichannel synthetic aperture radar images despeckling SAR images fast algorithms image amplitude radar reflectivity Bermudez-Moreno algorithm iterative shrinkage thresholding algorithm multipolarimetric images multitemporal RADARSAT-2 images visual quality evaluation indexes edge preservation

来源：评论

学校读者我要写书评

暂无评论

Two-Branch Deconvolutional Network With Application in Stereo Matching

引用

IEEE TRANSACTIONS ON IMAGE PROCESSING 2022年 31卷 327-340页

作者： Cheng, Chunbo Li, Hong Zhang, Liming Huazhong Univ Sci & Technol Sch Math & Stat Wuhan 430074 Peoples R China Hubei Polytech Univ Sch Math & Phys Huangshi 435003 Hubei Peoples R China Univ Macau Fac Sci & Technol Macau Peoples R China

Deconvolutional networks have attracted extensive attention and have been successfully applied in the field of computer vision. In this paper we propose a novel two-branch deconvolutional network (TBDN) that can improve the performance of conventional deconvolutional networks and reduce the computational complexity. A feasible iterative algorithm is designed to solve the optimization problem for the TBDN model, and a theoretical analysis of the convergence and computational complexity for the algorithm is also provided. The application of the TBDN in stereo matching is presented by constructing a disparity estimation network. Extensive experimental results on four commonly used datasets demonstrate the efficiency and effectiveness of the proposed TBDN.

关键词： Image reconstruction Feature extraction Costs Estimation Convolutional neural networks Convolution Computational modeling Two-branch deconvolutional network stereo matching iterative shrinkage thresholding algorithm disparity estimation network

来源：评论

学校读者我要写书评

暂无评论

Gradient-Based Methods for Sparse Recovery

引用

SIAM JOURNAL ON IMAGING SCIENCES 2011年第1期4卷 146-165页

作者： Hager, William W. Phan, Dzung T. Zhang, Hongchao Univ Florida Dept Math Gainesville FL 32611 USA IBM Corp Thomas J Watson Res Ctr Dept Business Analyt & Math Sci Yorktown Hts NY 10598 USA Louisiana State Univ Dept Math Ctr Computat & Technol Baton Rouge LA 70803 USA

The convergence rate is analyzed for the sparse reconstruction by separable approximation (SpaRSA) algorithm for minimizing a sum f(x) + psi(x), where f is smooth and psi is convex, but possibly nonsmooth. It is shown that if f is convex, then the error in the objective function at iteration k is bounded by a/k for some a independent of k. Moreover, if the objective function is strongly convex, then the convergence is R-linear. An improved version of the algorithm based on a cyclic version of the BB iteration and an adaptive line search is given. The performance of the algorithm is investigated using applications in the areas of signal processing and image reconstruction.

关键词： sparse reconstruction by separable approximation iterative shrinkage thresholding algorithm sparse recovery sublinear convergence linear convergence image reconstruction denoising compressed sensing nonsmooth optimization nonmonotone convergence BB method

来源：评论

学校读者我要写书评

暂无评论

Compressed LISTA Exploiting Toeplitz Structure

Compressed LISTA Exploiting Toeplitz Structure

引用

IEEE Radar Conference (RadarConf)

作者： Fu, Rong Huang, Tianyao Liu, Yimin Eldar, Yonina C. Tsinghua Univ Dept Elect Engn Beijing 100084 Peoples R China Technion Israel Inst Technol Dept Elect Engn IL-32000 Haifa Israel

ISBN: (纸本)9781728116792

iterative shrinkage thresholding algorithm (ISTA) has been widely applied to solve linear inverse problems. To increase the rate of convergence, Learned-ISTA (LISTA) adopts deep learning techniques to learn the optimal algorithm parameters like the mutual inhibition matrices and filter matrices, which significantly reduces the number of iterations. However, the size of the learned mutual inhibition matrix exhibits quadratic growth in the length of sparse signal, which restricts the applicability of LISTA in some large-scale problems. In many applications such as direction-of-arrival (DOA) estimation, the learned mutual inhibition matrix naturally has a Toeplitz structure. Here we exploit the Toeplitz structure and propose a convolutional network, namely LISTA-Toeplitz, to reduce the memory cost. Simulation results show that LISTA-Toeplitz outperforms traditional ISTA in convergence speed and achieves a level of accuracy comparable to LISTA in DOA simulation.

关键词： Linear Inverse Problem iterative shrinkage thresholding algorithm Learned-ISTA single-snapshot DOA estimation Toeplitz matrix convolutional network

来源：评论

学校读者我要写书评

暂无评论

Brain Activity Estimation using EEG-only Recordings Calibrated with joint EEG-fMRI Recordings using Compressive Sensing 13

Brain Activity Estimation using EEG-only Recordings Calibrat...

引用

13th International Conference on Sampling Theory and Applications (SampTA)

作者： Ataei, Ali Amini, Arash Ghazizadeh, Ali Sharif Univ Technol EE Dept Tehran Iran

ISBN: (纸本)9781728137414

Electroencephalogram (EEG) is a noninvasive, low-cost brain recording tool with high temporal but poor spatial resolution. In contrast, functional magnetic resonance imaging (fMRI) is a rather expensive brain recording tool with high spatial and poor temporal resolution. In this study, we aim at recovering the brain activity (source localization and activity-intensity) with high spatial resolution using only EEG recordings. Each EEG electrode records a linear combination of the activities of various parts of the brain. As a result, a multi-electrode EEG recording represents the brain activities via a linear mixing matrix. Due to distance attenuation, this matrix is almost sparse. Using simultaneous recordings of fMRI and EEG, we estimate the mixing matrix (calibration). Since Blood Oxygen Level Dependent (BOLD) signal of fMRI is a measure of energy used by active brain region, it has a quadratic relation with the electric potential waveform emitted from each fMRI volume pixel (voxel). Assuming uncorrelated time series from different regions, we reformulate the (underdetermined) forward problem as a linear problem and solve it using the Orthogonal Matching Pursuit (OMP) method. Besides the mixing matrix, the brain activities are often sparse spatially. Thus, we employ the estimated mixing matrix to extract the activity intensity of various brain regions from EEG recordings using iterative shrinkage thresholding algorithm (ISTA). We verify the proposed method on synthetic data. In particular, we divide the gray matter of the brain into 300 regions and assume a 30%-sparse measurement matrix, as well as 5% of regions to be active simultaneously. Simulations results show 88% accuracy in localizing the sources and and 66% accuracy in activity intensity estimation.

关键词： Electroencephalogram Blood Oxygen Level Dependency Voxels Gray matter Orthogonal Matching Pursuit iterative shrinkage thresholding algorithm

来源：评论

学校读者我要写书评

暂无评论

3D Trajectory Reconstruction under Smooth Camera Motion 8

3D Trajectory Reconstruction under Smooth Camera Motion

引用

8th International Conference on Robotics, Control and Automation (ICRCA)

作者： Li, Yongwei Beijing Univ Technol Fac Informat Technol Beijing Peoples R China

ISBN: (纸本)9798350344738;9798350344721

3D trajectory reconstruction is a crucial task in Autonomous Driving, Augmented Reality and Scientific Training. However, in critical configurations where the camera is in smooth motion, this task remains challenging due to the lack of depth information. In this paper, we study the 3D reconstruction of human motion from 2D projections, aiming to reconstruct 3D trajectories more accurately in the critical configuration(i.e. Scenes with smooth camera motions), which remedy the limitation of the existing methods. We proposed a new function to search the sparse solution in the discrete cosine transformation domain with l(1) optimization, in order to build a good reconstruct space for critical configurations, which transforms the underconstrained system into a convex optimization problem and solve it with a new algorithm for the first time as we known. By improving the Cross-Validation, we propose a reconstruction criterion to adaptively select coefficient weights beta and the number of basis vectors K for each point, which makes the reconstruction satisfy a good situation. Experiments on synthetic datasets made on Human3.6m and real datasets show that our method is superior to existing methods in 3D reconstruction under smooth camera motion.

关键词： 3D Reconstruction Human Motion Compressive Sensing iterative shrinkage thresholding algorithm l(1) optimization

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：