检索结果-内蒙古大学图书馆

Learning robust and discriminative low-rank representations for face recognition with occlusion

Pattern Recognition 2017年 66卷 129-143页

作者： Guangwei Gao Jian Yang Xiao-Yuan Jing Fumin Shen Wankou Yang Dong Yue Institute of Advanced Technology Nanjing University of Posts and Telecommunications Nanjing China Jiangsu Engineering Laboratory of Big Data Analysis and Control for active distribution network Nanjing University of Posts and Telecommunications Nanjing China Fujian Provincial Key Laboratory of Information Processing and Intelligent Control (Minjiang University) Fuzhou China School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China School of Automation Nanjing University of Posts and Telecommunications Nanjing China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu China School of Automation Southeast University Nanjing China

For robust face recognition tasks, we particularly focus on the ubiquitous scenarios where both training and testing images are corrupted due to occlusions. Previous low-rank based methods stacked each error image into a vector and then used L 1 or L 2 norm to measure the error matrix. However, in the stacking step, the structure information of the error image can be lost. Depart from the previous methods, in this paper, we propose a novel method by exploiting the low-rankness of both the data representation and each occlusion-induced error image simultaneously, by which the global structure of data together with the error images can be well captured. In order to learn more discriminative low-rank representations, we formulate our objective such that the learned representations are optimal for classification with the available supervised information and close to an ideal-code regularization term. With strong structure information preserving and discrimination capabilities, the learned robust and discriminative low-rank representation (RDLRR) works very well on face recognition problems, especially with face images corrupted by continuous occlusions. Together with a simple linear classifier, the proposed approach is shown to outperform several other state-of-the-art face recognition methods on databases with a variety of face variations.

关键词： Face recognition Low-rank matrix recovery Nuclear norm

来源：评论

学校读者我要写书评

暂无评论

Sensor-based Multi-Robot Coverage control with Spatial Separation in Unstructured Environments

Sensor-based Multi-Robot Coverage Control with Spatial Separ...

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Xinyi Wang Jiwen Xu Chuanxiang Gao Yizhou Chen Jihan Zhang Chenggang Wang Yulong Ding Ben M. Chen Department of Mechanical and Automation Engineering The Chinese University of Hong Kong Shatin N.T. Hong Kong China Department of Automation Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai Jiao Tong University Shanghai P. R. China Department of Control Science and Engineering Tongji University Shanghai China Frontiers Science Center for Intelligent Autonomous Systems Ministry of Education China

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Multi-robot systems have increasingly become instrumental in tackling coverage problems. However, the challenge of optimizing task efficiency without compromising task success still persists, particularly in expansive, unstructured scenarios with dense obstacles. This paper presents an innovative, decentralized Voronoi-based coverage control approach to reactively navigate these complexities while guaranteeing safety. This approach leverages the active sensing capabilities of multi-robot systems to supplement GIS (Geographic Information System), offering a more comprehensive and real-time understanding of environments like post-disaster. Based on point cloud data, which is inherently non-convex and unstructured, this method efficiently generates collision-free Voronoi regions using only local sensing information through spatial decomposition and spherical mirroring techniques. Then, deadlock-aware guided map integrated with a gradient-optimized, centroid Voronoi-based coverage control policy, is constructed to improve efficiency by avoiding exhaustive searches and local sensing pitfalls. The effectiveness of our algorithm has been validated through extensive numerical simulations in high-fidelity environments, demonstrating significant improvements in task success rate, coverage ratio, and task execution time compared with others.

关键词： Point cloud compression System recovery Robot sensing systems Sensors Safety Multi-robot systems Task analysis

来源：评论

学校读者我要写书评

暂无评论

FireMatch: A Semi-Supervised Video Fire Detection Network Based on Consistency and Distribution Alignment

arXiv

引用

arXiv 2023年

作者： Lin, Qinghua Li, Zuoyong Zeng, Kun Fan, Haoyi Li, Wei Zhou, Xiaoguang School of Computer Science and Mathematics Fujian University of Technology Fuzhou350118 China Fujian Provincial Key Laboratory of Information Processing and Intelligent Control College of Computer and Control Engineering Minjiang University Fuzhou350121 China School of Computer and Artificial Intelligence Zhengzhou University Zhengzhou450001 China School of Automation Beijing University of Post and Telecommunications Beijing100876 China School of Economics and Management Minjiang University Fuzhou350121 China

Deep learning techniques have greatly enhanced the performance of fire detection in videos. However, video-based fire detection models heavily rely on labeled data, and the process of data labeling is particularly costly and time-consuming, especially when dealing with videos. Considering the limited quantity of labeled video data, we propose a semi-supervised fire detection model called FireMatch, which is based on consistency regularization and adversarial distribution alignment. Specifically, we first combine consistency regularization with pseudo-label. For unlabeled data, we design video data augmentation to obtain corresponding weakly augmented and strongly augmented samples. The proposed model predicts weakly augmented samples and retains pseudo-label above a threshold, while training on strongly augmented samples to predict these pseudo-labels for learning more robust feature representations. Secondly, we generate video cross-set augmented samples by adversarial distribution alignment to expand the training data and alleviate the decline in classification performance caused by insufficient labeled data. Finally, we introduce a fairness loss to help the model produce diverse predictions for input samples, thereby addressing the issue of high confidence with the non-fire class in fire classification scenarios. The FireMatch achieved an accuracy of 76.92% and 91.81% on two real-world fire datasets, respectively. The experimental results demonstrate that the proposed method outperforms the current state-of-the-art semi-supervised classification methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Fires

来源：评论

学校读者我要写书评

暂无评论

Sparse View HRRP Recognition Based on Dual-Task of Generation and Recognition Method

Sparse View HRRP Recognition Based on Dual-Task of Generatio...

引用

Signal, Information and Data processing (ICSIDP), IEEE International Conference on

作者： Yuanhao Meng Lue Wang Qiang Zhou Xin Zhang Liang Zhang Yanhua Wang Radar Technology Research Institute Beijing Institute of Technology Beijing China Xi'an Modern Control Technology Research Institute Xi'an China Zhengzhou Academy of Intelligent Technology Beijing Institute of Technology Zhengzhou China Beijing Key Laboratory of Embedded Real-Time Information: Processing Technology Beijing China Chongqing Innovation Center Beijing Institute of Technology China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

In recent years, high-resolution range profiles (HRRPs) have played a significant role in radar target recognition. The pose sensitivity of HRRPs results in unsatisfactory target recognition capabilities under sparse viewing angles. However, for non-cooperative targets, obtaining full-view HRRPs is often challenging, leading to suboptimal performance in HRRP-based target recognition. To address the issue of target recognition under sparse viewing angles, we propose an end-to-end model DGR for sparse view recognition that integrates combined generation and recognition tasks. The DGR model extracts deep features of HRRPs through an attention mechanism and utilizes these features for both target recognition and the generation of adjacent-view HRRPs, with both tasks mutually reinforcing each other for improved performance in recognition and data generation. Furthermore, by employing domain adaptation methods, the model learns the variability of target HRRPs in the source domain and applies this knowledge to the target domain, enhancing its generalization ability. Experimental results demonstrate that the DGR model achieves optimal recognition and generation capabilities simultaneously.

关键词： Adaptation models Sensitivity Target recognition Radar Data collection Feature extraction Generative adversarial networks Data augmentation Data models Data mining

来源：评论

学校读者我要写书评

暂无评论

EEG-Based Brain-Computer Interfaces Are Vulnerable to Backdoor Attacks

Research Square

引用

Research Square 2021年

作者： Meng, Lubin Huang, Jian Zeng, Zhigang Jiang, Xue Yu, Shan Jung, Tzyy-Ping Lin, Chin-Teng Chavarriaga, Ricardo Wu, Dongrui Ministry of Education Key Laboratory of Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China Brainnetome Center and National Laboratory of Pattern Recognition Institute of Automation Chinese Academy of Sciences Beijing China La Jolla CA United States Center for Advanced Neurological Engineering Institute of Engineering in Medicine UCSD La Jolla CA United States Centre of Artificial Intelligence Faculty of Engineering and Information Technology University of Technology Sydney Australia ZHAW DataLab Zürich University of Applied Sciences Winterthur8401 Switzerland

Research and development of electroencephalogram (EEG) based brain-computer interfaces (BCIs) have advanced rapidly, partly due to the wide adoption of sophisticated machine learning approaches for decoding the EEG signals. However, recent studies have shown that machine learning algorithms are vulnerable to adversarial attacks, e.g., the attacker can add tiny adversarial perturbations to a test sample to fool the model, or poison the training data to insert a secret backdoor. Previous research has shown that adversarial attacks are also possible for EEG-based BCIs. However, only adversarial perturbations have been considered, and the approaches are theoretically sound but very difficult to implement in practice. This article proposes to use narrow period pulse for poisoning attack of EEG-based BCIs, which is more feasible in practice and has never been considered before. One can create dangerous backdoors in the machine learning model by injecting poisoning samples into the training set. Test samples with the backdoor key will then be classified into the target class specified by the attacker. What most distinguishes our approach from previous ones is that the backdoor key does not need to be synchronized with the EEG trials, making it very easy to implement. The effectiveness and robustness of the backdoor attack approach is demonstrated, highlighting a critical security concern for EEG-based BCIs. © 2021, CC BY.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

An Improved Lmmse Channel Estimation Algorithm for Mimo-Ofdm System

SSRN

引用

SSRN 2022年

作者： Tang, Wanwei Wang, Lixia Wang, Zedai Wei, Mingzhe Zhao, Jinlong Yue, Xinwei The Key Laboratory of Intelligent Data Information Processing and Control of Hebei Province Tangshan University Tangshan063000 China The School of Engineering Hong Kong University of Science and Technology 999077 Hong Kong The School of Electronic and Information Engineering Beihang University Beijing100191 China The School of Information and Communication Engineering Beijing Information Science and Technology University Beijing100101 China

This paper studies channel estimation techniques to recover the received signal more accurately and quickly. The classical Least Squares (LS) and Linear Minimum Mean-squared Error (LMMSE) estimation have the problems of poor estimation performance and high complexity. An Fast-LMMSE channel estimation algorithm is proposed to balance the performance and complexity. In order to apply it to the actual system, the corresponding pilot is also designed. Simulation results are provided and the comparison shows the performance gains achieved by the proposed technique over LS and MMSE estimation in SISO and MIMO communication systems. © 2022, The Authors. All rights reserved.

关键词： Channel estimation

来源：评论

学校读者我要写书评

暂无评论

End-to-end Driving in High-Interaction Traffic Scenarios with Reinforcement Learning

arXiv

引用

arXiv 2024年

作者： Li, Yueyuan Jiang, Mingyang Zhang, Songan Yuan, Wei Wang, Chunxiang Yang, Ming The Department of Automation Shanghai Jiao Tong University Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China The Global Institute of Future Technology Shanghai Jiao Tong University Shanghai200240 China The Innovation Center of Intelligent Connected Vehicles Global Institute of Future Technology Shanghai Jiao Tong University Shanghai200240 China

Dynamic and interactive traffic scenarios pose significant challenges for autonomous driving systems. Reinforcement learning (RL) offers a promising approach by enabling the exploration of driving policies beyond the constraints of pre-collected datasets and predefined conditions, particularly in complex environments. However, a critical challenge lies in effectively extracting spatial and temporal features from sequences of high-dimensional, multi-modal observations while minimizing the accumulation of errors over time. Additionally, efficiently guiding large-scale RL models to converge on optimal driving policies without frequent failures during the training process remains tricky. We propose an end-to-end model-based RL algorithm named Ramble to address these issues. Ramble processes multi-view RGB images and LiDAR point clouds into low-dimensional latent features to capture the context of traffic scenarios at each time step. A transformer-based architecture is then employed to model temporal dependencies and predict future states. By learning a dynamics model of the environment, Ramble can foresee upcoming traffic events and make more informed, strategic decisions. Our implementation demonstrates that prior experience in feature extraction and decision-making plays a pivotal role in accelerating the convergence of RL models toward optimal driving policies. Ramble achieves state-of-the-art performance regarding route completion rate and driving score on the CARLA Leaderboard 2.0, showcasing its effectiveness in managing complex and dynamic traffic situations. Copyright © 2024, The Authors. All rights reserved.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Determination of Pattern Recognition Problems based on a Pythagorean Fuzzy Correlation Measure from Statistical Viewpoint

Determination of Pattern Recognition Problems based on a Pyt...

引用

International Workshop on Advanced Computational Intelligence (IWACI)

作者： Paul Augustine Ejegwa Yuming Feng Shiping Wen Wei Zhang Key Laboratory of Intelligent Information Processing and Control Chongqing Three Gorges University Wanzhou Chongqing China School of Computer Science and Engineering Chongqing Three Gorges University Wanzhou Chongqing China Australian AI Institute University of Technology Sydney Ultimo Australia School of Three Gorges Artificial Intelligence Chongqing Three Gorges University Wanzhou Chongqing China

By considering all the three parameters describing Pythagorean fuzzy set, we introduce a new technique of computing correlation coefficient between Pythagorean fuzzy sets from statistical perspective. The correlation coefficient value obtain via this technique shows strength of correlation between the Pythagorean fuzzy sets and indicates whether the Pythagorean fuzzy sets under consideration are related negatively or positively in contrast to other existing correlation coefficient approaches in Pythagorean fuzzy context, which only assess the strength of relationship. Certain numerical examples are considered to ascertain the authenticity of this method over similar techniques studied in intuitionistic/Pythagorean fuzzy contexts. Some pattern recognition problems are resolved with the aid of the new technique. For higher productivity sake, this technique could be approached from an object-oriented perspective.

关键词： Correlation coefficient Productivity Fuzzy sets Correlation Uncertainty Pattern recognition Performance analysis

来源：评论

学校读者我要写书评

暂无评论

Event-Triggered Model Predictive control for A Three-Phase Inverter with Output LC Filter

Event-Triggered Model Predictive Control for A Three-Phase I...

引用

IEEE Conference on Industrial Electronics and Applications (ICIEA)

作者： Benfei WANG Jingjing HUANG Jose RODRIGUEZ Cristian GARCIA Guodong FENG School of Intelligent Systems Engineering Sun Yat-sen University Guangdong China Key Laboratory of System Control and Information Processing Ministry of Education Shanghai China School of Electronic and Information Engineering Xi’an Jiaotong University Xi’an China Engineering School Universidad Andrés Bello Santiago Chile Engineering Faculty Universidad de Talca Curico Chile

ISBN: (数字)9781728151694

ISBN: (纸本)9781728151700

This paper proposes an event-triggered model predictive control (ET-MPC) method for the three-phase inverter with output LC filter. The proposed method is developed by combining the conventional finite-control-set MPC (FCS-MPC) method and the ET control strategy. The ET control strategy incorporates a preset triggering condition based on the requirement of the regulation performance for the three-phase inverter. When the state of the three-phase inverter activates the triggering condition, FCS-MPC is triggered with the updated state so that the optimal control actions can be generated when it is necessary. If the state of the inverter is within the preset triggering condition, FCS-MPC is suspended and the control actions are held as constant. Consequently, the unnecessary online prediction, online optimization, and the corresponding control actions can be avoided. Therefore, with the tolerate regulation performance, the proposed ET-MPC can reduce the computational burden of the conventional FCS-MPC. The various simulation studies of the three-phase inverter are conducted and discussed to demonstrate the effectiveness of the proposed ET-MPC method.

关键词： Computational modeling Simulation Switches Inverters Regulation Optimization Predictive control

来源：评论

学校读者我要写书评

暂无评论

Open Set HRRP Recognition Using Adaptive Learning with Enclosed Decision Boundaries

Open Set HRRP Recognition Using Adaptive Learning with Enclo...

引用

Signal, Information and Data processing (ICSIDP), IEEE International Conference on

作者： Binhua Yu Ping-An Liu Yichen Liu Hongyu Wang Liang Zhang Yanhua Wang Kun Liu Radar Technology Research Institute Beijing Institute of Technology Beijing China Xi’an Modern Control Technology Research Institute Xi’an China Zhengzhou Academy of Intelligent Technology Beijing Institute of Technology Zhengzhou China Beijing Key Laboratory of Embedded Real-Time Information: Processing Technology Beijing China Chongqing Innovation Center Beijing Institute of Technology Chongqing China

ISBN: (数字)9798331515669

ISBN: (纸本)9798331515676

High resolution range profiles (HRRP), which contain fine geometric structure information of targets, plays animportant role in the field of Radar automatic target recognition (RATR). In the actual application of RATR, radar may often capture targets that are non-cooperative, making it difficult to establish a comprehensive database, and traditional recognition algorithms are prone to misjudgment when facing unknown targets. Therefore, Open set recognition (OSR) is proposed to simultaneously recognize in-database and out-of-database targets. Considering the pose sensitivity of HRRP, minor changes in the target’s pose can cause significant changes in the HRRP images, leading to a potentially dispersed distribution of samples from the same category in the feature space, which makes it difficult for traditional fixed decision boundaries in OSR to adapt. To address this, this paper proposes an HRRP open set recognition method capable of adaptively learning closed decision boundaries. Based on Hierarchical dirichlet processes (HDP), the method achieves adaptive clustering of subclasses under different categories according to the data distribution characteristics of each category, determining multiple small decision boundaries within each category, and allowing the decision boundaries to dynamically adjust according to the sample features. Finally, experiments with real HRRP data have verified the effectiveness of the proposed method.

关键词： Training Sensitivity image resolution image recognition Target recognition Databases Imaging Radar applications Signal resolution Dispersion

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：