Article
Engineering, Electrical & Electronic
Junlin Li, Wei Zhou, Xiuting Li
Summary: In this paper, we study a broad class of nonconvex and nonsmooth composition optimization problems and propose a proximal alternating partially linearized minimization (PAPLM) algorithm to solve them. The effectiveness and superior performance of the proposed algorithm are demonstrated through theoretical analysis and numerical experiments.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2023)
Article
Computer Science, Hardware & Architecture
Chaofan Wang, Yuxin Zhang, Liying Sun, Jiefei Han, Lianying Chao, Lisong Yan
Summary: This paper proposes a variable step size sparsity adaptive matching pursuit (SAMPVSS) algorithm, which constructs a set of candidate atoms by calculating the correlation between the measurement matrix and the residual and selects the atom most related to the residual. The algorithm introduces an exponential function to determine the number of atoms to be selected each time and sets different step sizes based on the iteration stage. Simulation results show that the proposed algorithm has good reconstruction effects on both one-dimensional and two-dimensional signals.
Article
Computer Science, Information Systems
Jun Wang
Summary: In this paper, a wonderful triangle is introduced to explore the concrete metric relationship between llxll1/llxll and llxll0. Based on the analysis of the iterative soft-thresholding operator, the angle of the triangle corresponding to the side llxll. - llxll1/llxll0 is studied, demonstrating the meaningfulness of signal sparsity within a certain exact interval.
INFORMATION SCIENCES
(2022)
Article
Mathematics, Applied
Hui Zhang, Lu Zhang, Hao-Xing Yang
Summary: The linearized Bregman iterations (LBreI) and its variants have attracted considerable attention in signal/image processing and compressed sensing. However, the assumption of Lipschitz gradient continuity restricts their practical applications. This study proposes a generalized algorithmic framework and discovers that the Lipschitz gradient continuity assumption can be replaced by a Lipschitz-like convexity condition in both convex and nonconvex cases. Additionally, the proposed framework can solve a class of bilevel optimization problems, extending the previous research by Cai et al.
MATHEMATICS OF COMPUTATION
(2023)
Article
Engineering, Electrical & Electronic
Xiaotao Shao, Caike Wei, Yi Xie, Zhongli Wang, Yan Shen
Summary: The proposed compressed sensing MRI method with shearlet dictionary and non-local similarity model can reconstruct MRI images with less data and higher quality compared to traditional methods.
IET SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Yuhan Li, Tianyao Huang, Xingyu Xu, Yimin Liu, Lei Wang, Yonina C. Eldar
Summary: This study investigates the phase transitions of range-Doppler recovery in FAR using compressed sensing methods. The results show that block sparse recovery outperforms standard recovery when extended targets occupy multiple range cells, facilitating radar parameter design.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Chemistry, Multidisciplinary
Xue Bi, Lu Leng, Cheonshik Kim, Xinwen Liu, Yajun Du, Feng Liu
Summary: The paper proposes a constrained backtracking matching pursuit (CBMP) algorithm for image reconstruction, which effectively controls the increment of the estimated sparsity level and accurately estimates the true support set, with an energy criterion and an improved 4:1 rule as constraints, achieving better performance and further stability.
APPLIED SCIENCES-BASEL
(2021)
Article
Radiology, Nuclear Medicine & Medical Imaging
Kirsten Koolstra, Rob Remis
Summary: Learning a trainable preconditioner can significantly accelerate parallel imaging and compressed sensing reconstructions, achieving similar performance to state-of-the-art preconditioners. The method is generalizable to different sampling schemes, coil configurations, and anatomies.
MAGNETIC RESONANCE IN MEDICINE
(2022)
Article
Geochemistry & Geophysics
Bangjie Zhang, Gang Xu, Hanwen Yu, Hui Wang, Hao Pei, Wei Hong
Summary: This article proposes a novel robust gridless compressed sensing (RGLCS) algorithm for high-resolution 3-D imaging. The algorithm uses atomic norm minimization to model the joint-sparsity pattern on elevation distribution between adjacent pixels, and models outliers and disturbances as sparsely distributed spike noise in the image domain. Experimental results validate the effectiveness of the proposed algorithm.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Engineering, Electrical & Electronic
Sanhita Guha, Andreas Bathelt, Miguel Heredia Conde, Joachim Ender
Summary: Existing compressed sensing algorithms fail in radar target detection when there is a large gap in the frequency band. A new algorithm based on a subdivision-fusion scheme is proposed to solve this problem. The algorithm utilizes a structured sensing matrix based on radar signals to achieve good range resolution despite high coherence. The performance of the algorithm is discussed and demonstrated through simulative examples and real measurement data, showing superior performance in the presence of band gaps.
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING
(2023)
Article
Automation & Control Systems
Amir Moslemi
Summary: Compressive sensing is applied to reduce the number of samples required for classification in representation learning. A novel approach is presented where image pixels are treated as sensors to identify the optimal sensors in feature space. Spatial sensor locations are learned to identify discriminative information regions within images. L1-2 minimization, RRQR and SVM are used for sparse minimization, feature space extraction and discrimination vector acquiring. The proposed method is evaluated on four experiments and outperforms a state-of-art technique.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2023)
Article
Automation & Control Systems
Yunyi Li, Yiqiu Jiang, Hengmin Zhang, Jianxun Liu, Xiangling Ding, Guan Gui
Summary: Compressive sensing (CS) based computed tomography (CT) image reconstruction aims to reduce radiation risk by using sparse-view projection data. However, it is challenging to achieve satisfactory image quality from incomplete projections. This paper proposes an L 1 / 2-regularized nonlocal self-similarity (NSS) denoiser based CT reconstruction model that combines with low-rank approximation and group sparse coding (GSC) framework. Experimental results on clinical CT images show that the proposed approach outperforms popular approaches.
JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS
(2023)
Article
Engineering, Electrical & Electronic
Sujit Das, Jyotsna Kumar Mandal
Summary: The theory of compressed sensing suggests that signals can be recovered from fewer samples by solving a minimization problem. However, using randomized sensing matrices for large-scale 2D image signals consumes significant computational resources. This paper proposes a new technique for better reconstruction performance using sparsity normalization and non-adaptive measurements.
SIGNAL IMAGE AND VIDEO PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Guangxiang Li, Shidong Li, Dequan Li, Chi Ma
Summary: A tail-Hadamard product parametrization (tail-HPP) approach is proposed for sparse signal recovery in compressed sensing. The algorithm combines the efficiency of the HPP technique and the greater capacity of signal recovery enabled by the tail- t 1 -minimization approach. It is proven that the tail-HPP approach is equivalent to the tail- t 1 -minimization problem. Extensive simulation experiments confirm the superiority of the tail-HPP algorithm compared to state-of-the-art sparse recovery techniques.
Article
Engineering, Biomedical
Tariq Rahim, Ledya Novamizanti, I. Nyoman Apraz Ramatryana, Soo Young Shin
Summary: Efficient image sampling and transfer are crucial in medical imaging research. The compressed sensing theory demonstrates the possibility of compression during data retrieval, leading to high reconstruction image quality in various medical images using the proposed multiple basis reweighted analysis (M-BRA) method.
COMPUTERIZED MEDICAL IMAGING AND GRAPHICS
(2021)
Article
Operations Research & Management Science
Tao Sun, Yuejiao Sun, Yangyang Xu, Wotao Yin
COMPUTATIONAL OPTIMIZATION AND APPLICATIONS
(2020)
Article
Computer Science, Artificial Intelligence
Chun-Na Li, Yuan-Hai Shao, Wotao Yin, Ming-Zeng Liu
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2020)
Article
Mathematics, Applied
Xinmeng Huang, Ernest K. Rya, Wotao Yin
JOURNAL OF MATHEMATICAL ANALYSIS AND APPLICATIONS
(2020)
Article
Mathematics, Applied
Jialin Liu, Wotao Yin, Wuchen Li, Yat Tin Chow
Summary: The paper proposes a fast algorithm for calculating the Wasserstein-1 distance, based on multilevel primal-dual algorithms, and demonstrates its computational speed through numerical examples and complexity analysis. The proposed algorithm provides solutions within 0.2 to 1.5 seconds on a single CPU for commonly used image examples of size 512 x 512, which is much faster than state-of-the-art algorithms.
SIAM JOURNAL ON SCIENTIFIC COMPUTING
(2021)
Article
Computer Science, Software Engineering
Ernest K. Ryu, Robert Hannah, Wotao Yin
Summary: This paper introduces a geometric approach to analyzing the convergence of fixed point iterations by utilizing a new tool called the scaled relative graph. This tool establishes a correspondence between nonlinear operators and subsets of the 2D plane, allowing for a rigorous proof of convergence through geometric arguments in the 2D plane.
MATHEMATICAL PROGRAMMING
(2022)
Article
Mathematics, Applied
Yiyang Liu, Zaiwen Wen, Wotao Yin
Summary: The goal of this paper is to efficiently solve the large-scale linear programming formulation of Optimal Transport problems. The key observations are the sparsity of primal solutions and the good geometric properties of the cost function. Based on these observations, the paper proposes an algorithm that utilizes a hierarchical multiscale structure to solve large-scale OT problems and significantly improve efficiency in the computation process.
JOURNAL OF SCIENTIFIC COMPUTING
(2022)
Article
Mathematics, Applied
Jinshan Zeng, Wotao Yin, Ding-Xuan Zhou
Summary: The augmented Lagrangian method (ALM) is a useful method for constrained optimization, but it can experience oscillations and divergence when the underlying problem is nonconvex and nonsmooth. This paper modifies ALM to use a Moreau envelope and establishes its convergence. Two practical variants are also proposed.
JOURNAL OF SCIENTIFIC COMPUTING
(2022)
Article
Mathematics, Applied
Howard Heaton, Samy Wu Fung, Alex Tong Lin, Stanley Osher, Wotao Yin
Summary: Inverse problems are important in recovering signals from noisy measurements. This study proposes a new algorithm called Wasserstein-based projection (WP) that approximates the true projection with high probability, providing theoretical guarantees for optimization methods in signal recovery.
SIAM JOURNAL ON MATHEMATICS OF DATA SCIENCE
(2022)
Article
Engineering, Electrical & Electronic
Tianyi Chen, Yuejiao Sun, Wotao Yin
Summary: Stochastic compositional optimization generalizes classic stochastic optimization for minimizing compositions of functions, with applications in reinforcement learning and meta learning. The new Stochastically Corrected Stochastic Compositional gradient method (SCSC) ensures convergence at the same rate as traditional methods and can be accelerated with SGD techniques. Applying Adam to SCSC achieves state-of-the-art performance in stochastic compositional optimization, tested in model-agnostic meta-learning tasks.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Tianyi Chen, Yuejiao Sun, Wotao Yin
Summary: The algorithm LASG is developed to solve distributed learning problems efficiently, saving communication costs and tailored for stochastic gradients. Through introducing new rules and analysis, LASG achieves impressive empirical performance in practice.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Huan Li, Cong Fang, Wotao Yin, Zhouchen Lin
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2020)
Article
Mathematics, Applied
Omri Azencot, Wotao Yin, Andrea Bertozzi
SIAM JOURNAL ON APPLIED DYNAMICAL SYSTEMS
(2019)
Article
Engineering, Electrical & Electronic
Xinwei Zhang, Mingyi Hong, Sairaj Dhople, Wotao Yin, Yang Liu
Summary: This paper studies the behavior of the FedAvg algorithm in Federated Learning (FL) and proposes a new algorithm design strategy to design FL algorithms that are fast and require minimal assumptions, achieving optimal optimization and communication complexity, accommodating various local computation models. The new algorithms are communication efficient, with communication effort reducing as the heterogeneity among local data decreases.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Yanli Liu, Yuejiao Sun, Wotao Yin
Summary: This paper develops new algorithms for decentralized machine learning over a network to reduce computation and communication complexities by running warm-started Katyusha algorithm. Experimental results demonstrate that these algorithms can significantly reduce computational and communication costs compared to state-of-the-art techniques.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2021)
Article
Engineering, Electrical & Electronic
Xianghui Mao, Kun Yuan, Yubin Hu, Yuantao Gu, Ali H. Sayed, Wotao Yin
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2020)