检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Xie, Tao Dai, Kun Wang, Ke Li, Ruifeng Zhao, Lijun State Key Laboratory of Robotics and System Harbin Institute of Technology Harbin150006 China SenseTime Group Inc. Beijing100080 China

Local feature matching between images remains a challenging task, especially in the presence of significant appearance variations, e.g., extreme viewpoint changes. In this work, we propose DeepMatcher, a deep Transformer-based network built upon our investigation of local feature matching in detector-free methods. The key insight is that local feature matcher with deep layers can capture more human-intuitive and simpler-to-match features. Based on this, we propose a Slimming Transformer (SlimFormer) dedicated for DeepMatcher, which leverages vector-based attention to model relevance among all keypoints and achieves long-range context aggregation in an efficient and effective manner. A relative position encoding is applied to each SlimFormer so as to explicitly disclose relative distance information, further improving the representation of keypoints. A layer-scale strategy is also employed in each SlimFormer to enable the network to assimilate message exchange from the residual block adaptively, thus allowing it to simulate the human behaviour that humans can acquire different matching cues each time they scan an image pair. To facilitate a better adaption of the SlimFormer, we introduce a Feature Transition Module (FTM) to ensure a smooth transition in feature scopes with different receptive fields. By interleaving the self- and cross-SlimFormer multiple times, DeepMatcher can easily establish pixel-wise dense matches at coarse level. Finally, we perceive the match refinement as a combination of classification and regression problems and design Fine Matches Module to predict confidence and offset concurrently, thereby generating robust and accurate matches. Experimentally, we show that DeepMatcher significantly outperforms the state-of-the-art methods on several benchmarks, demonstrating the superior matching capability of DeepMatcher. The code is available at https://***/XT1997/DeepMatcher. © 2023, CC BY.

关键词： Behavioral research

来源：评论

学校读者我要写书评

暂无评论

Diverse Gaussian Sampling for Human Motion Prediction 16th

Diverse Gaussian Sampling for Human Motion Prediction

引用

16th International Conference on Social robotics, ICSR + InnoBiz 2024

作者： Luo, Jiefu Wang, Jiansheng Liu, Zhenfei Cheng, Jun Guangdong Provincial Key Laboratory of Robotics and Intelligent System Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China University of Chinese Academy of Sciences Beijing China The Chinese University of Hong Kong Hong Kong Department of Pediatric Orthopedics Shenzhen Children’s Hospital Guangdong Shenzhen518000 China

ISBN: (纸本)9789819611508

In many industrial applications, it is necessary to obtain accurate human motion prediction. The prediction of human movement based on a 3D skeleton, which forecasts future postures using the human skeletal structure, significantly aids in enabling machines to understand human behavior and react more intelligently. Additionally, by predicting human actions, potential dangers can be perceived, especially in safety-related issues, necessitating a broader range of predictions. In this paper, considering the greater diversity generated by Variational Autoencoders (VAEs), we employ VAEs for prediction. In contrast to other studies, our primary focus is on obtaining more diverse samples. Extensive experiments demonstrate that our proposed model performs well on the Human 3.6M [35] and HumanEva-I [36] datasets. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Improved Capon Estimator for High-Resolution DOA Estimation and Its Statistical Analysis

引用

IEEE/CAA Journal of Automatica Sinica 2023年第8期10卷 1716-1729页

作者： Weiliang Zuo Jingmin Xin Changnong Liu Nanning Zheng Akira Sano IEEE the National Key Laboratory of Human-Machine Hybrid Augmented Intelligence the National Engineering Research Center for Visual Information and Applicationsand the Institute of Artificial Intelligence and RoboticsXi’an Jiaotong UniversityXi’an 710049China the Data Center China Construction BankBeijing 100033China the Department of System Design Engineering Keio UniversityYokohama 223-8522Japan

Despite some efforts and attempts have been made to improve the direction-of-arrival(DOA)estimation performance of the standard Capon beamformer(SCB)in array processing,rigorous statistical performance analyses of these modified Capon estimators are still *** paper studies an improved Capon estimator(ICE)for estimating the DOAs of multiple uncorrelated narrowband signals,where the higherorder inverse(sample)array covariance matrix is used in the Capon-like cost *** establishing the relationship between this nonparametric estimator and the parametric and classic subspace-based MUSIC(multiple signal classification),it is clarified that as long as the power order of the inverse covariance matrix is increased to reduce the influence of signal subspace components in the ICE,the estimation performance of the ICE becomes equivalent to that of the MUSIC regardless of the signal-to-noise ratio(SNR).Furthermore the statistical performance of the ICE is analyzed,and the large-sample mean-squared-error(MSE)expression of the estimated DOA is *** the effectiveness and the theoretical analysis of the ICE are substantiated through numerical examples,where the Cramer-Rao lower bound(CRB)is used to evaluate the validity of the derived asymptotic MSE expression.

关键词： Capon beamformer direction-of-arrival(DOA)estimation large-sample mean-squared-error(MSE) subspace-based methods uniform linear array

来源：评论

学校读者我要写书评

暂无评论

Model heat source using actual distribution of laser power density for simulation of laser processing

引用

Journal of Central South University 2022年第10期29卷 3277-3293页

作者： WANG Gen-wang DING Ye GUAN Yan-chao WANG Yang YANG Li-jun State Key Laboratory of Robotics and Systems Harbin Institute of TechnologyHarbin 150001China School of Mechatronics Engineering Harbin Institute of TechnologyHarbin 150001China

The model of heat source(MHS) which reflects the thermal interaction between materials and laser during processing determines the accuracy of simulation results. To acquire desirable simulations results, although various modifications of heat sources in the aspect of absorption process of laser by materials have been purposed, the distribution of laser power density(DLPD) in MHS is still modeled theoretically. However, in the actual situations of laser processing, the DLPD is definitely different from the ideal models. So, it is indispensable to build MHS using actual DLPD to improve the accuracy of simulation results. Besides, an automatic modeling method will be benefit to simplify the tedious pre-processing of simulations. This paper presents a modeling method and corresponding algorithm to model heat source using measured DLPD. This algorithm automatically processes original data to get modeling parameters and provides a step MHS combining with absorption models. Simulations and experiments of heat transfer in steel plates irradiated by laser prove the mothed and the step MHS. Moreover, the investigations of laser induced thermal-crack propagation in glass highlight the signification of modeling heat source based on actual DLPD and demonstrate the enormous application of this method in the simulation of laser processing.

关键词： heat source laser processing distribution of power density digital images processing heat transfer

来源：评论

学校读者我要写书评

暂无评论

A Single Image Deraining Network Based on Global Feature Perception 22

A Single Image Deraining Network Based on Global Feature Per...

引用

Proceedings of the 2022 6th International Conference on Computer Science and Artificial Intelligence

作者： Bo Fu Hongguang Wang State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences China and University of Chinese Academy of Sciences China State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences China and Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences China

ISBN: (纸本)9781450397773

Rainy weather can degrade the quality of image captured outdoors, which in turn reduces the effectiveness of the subsequent computer vision algorithm. Therefore, as a way to improve the performance of subsequent visual tasks, single image deraining has become an important research topic. In this paper, a single image deraining method based on global feature perception is proposed, which can effectively locate the rain streak and estimate corresponding background content via Swin Transformer mechanism. Different from convolutional neural network and traditional Transformer, the proposed network can extract the global features of an image while reducing the amount of computation. Combing advantages of Swin Transformer and convolutional network, a three-stage deraining network is proposed. First, the coarse features of an image are extracted through the convolutional layer. Then, using the feature refining module based on Swin Transformer, global features are extracted and fused. Finally, through the image reconstruction module, the clean image is generated by using residual learning. The experimental results show the superiority of the proposed network.

关键词： single image deraining Transformer convolution neural network

来源：评论

学校读者我要写书评

暂无评论

A High-Accuracy Slotted Hole Detector 4

A High-Accuracy Slotted Hole Detector

引用

4th International Conference on robotics and Computer Vision, ICRCV 2022

作者： Tao, Liming Xia, Renbo Zhao, Jibin Li, Yinghao Zou, Hangbo Wang, Fangyuan State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences Shenyang110016 China Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang110169 China University of Chinese Academy of Sciences Beijing100049 China

ISBN: (数字)9781665481700

ISBN: (纸本)9781665481700

Slotted holes are often integrated into industrial parts to flexibly adjust the position of the parts. The measurement of the geometric parameters of the slotted hole is the key to ensure the mechanical properties of the parts. Identifying and detecting the edges of slotted holes on the image is the basis for the slotted hole measurement with high-accuracy. This is a challenging problem because the projection of the slotted hole on the image is not a primitive geometry. In this paper, a high-accuracy slotted hole detector is designed. First, a curvature-based algorithm is introduced to detect the dominant points on the curves of interest after edge detection with Canny detector. Then, a dominant point verification algorithm based on a new projective invariant is proposed to significantly remove incorrect dominant points on the edge of the slotted holes. Finally, a dominant point tuning algorithm based on line and ellipse fitting is developed for accurately finding the segmentations between lines and ellipse arcs to calculate slotted hole parameters. Extensive experiments on real images and synthetic images demonstrate that the proposed slotted hole detector has high detection accuracy and strong robustness. © 2022 IEEE.

关键词： Parameter estimation

来源：评论

学校读者我要写书评

暂无评论

OAMatcher: An Overlapping Areas-based Network for Accurate Local Feature Matching

arXiv

引用

arXiv 2023年

作者： Dai, Kun Xie, Tao Wang, Ke Jiang, Zhiqiang Li, Ruifeng Zhao, Lijun State Key Laboratory of Robotics and System Harbin Institute of Technology Harbin150006 China SenseTime Group Inc. Beijing100080 China

Local feature matching is an essential component in many visual applications. In this work, we propose OAMatcher, a Tranformer-based detector-free method that imitates humans behavior to generate dense and accurate matches. Firstly, OAMatcher predicts overlapping areas to promote effective and clean global context aggregation, with the key insight that humans focus on the overlapping areas instead of the entire images after multiple observations when matching keypoints in image pairs. Technically, we first perform global information integration across all keypoints to imitate the humans behavior of observing the entire images at the beginning of feature matching. Then, we propose Overlapping Areas Prediction Module (OAPM) to capture the keypoints in co-visible regions and conduct feature enhancement among them to simulate that humans transit the focus regions from the entire images to overlapping regions, hence realizeing effective information exchange without the interference coming from the keypoints in non-overlapping areas. Besides, since humans tend to leverage probability to determine whether the match labels are correct or not, we propose a Match Labels Weight Strategy (MLWS) to generate the coefficients used to appraise the reliability of the ground-truth match labels, while alleviating the influence of measurement noise coming from the data. Moreover, we integrate depth-wise convolution into Tranformer encoder layers to ensure OAMatcher extracts local and global feature representation concurrently. Comprehensive experiments demonstrate that OAMatcher outperforms the state-of-the-art methods on several benchmarks, while exhibiting excellent robustness to extreme appearance variants. The source code is available at https://***/DK-HU/OAMatcher. © 2023, CC BY.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Riemannian Geometry Based Frequency Band Optimization in Motor Imagery Classification

Riemannian Geometry Based Frequency Band Optimization in Mot...

引用

IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent systems

作者： Fanbo Zhuo Fengzhen Tang The State Key Laboratory of Robotics Shenyang Institute of Automation Shenyang China The Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang China The University of Chinese Academy of Sciences Beijing China

ISBN: (数字)9798331506056

ISBN: (纸本)9798331506063

This paper presents an approach for optimizing frequency bands in motor imagery (MI) classification using Particle Swarm Optimization (PSO) combined with the Generalized learning Riemannian space quantization (GLRSQ) classifier. The proposed method leverages the class distance metric to assess the distribution of EEG signals in the Symmetric Positive Definite (SPD) manifold. Experimental results on the BCI IIa dataset demonstrate improved classification accuracy and Kappa values compared to baseline method. The study highlights the effectiveness of opti-mizing frequency bands on the Riemannian manifold for enhancing MI classification performance.

关键词： Manifolds Measurement Training Time-frequency analysis Automatic frequency control Accuracy Quantization (signal) Optimization methods Motors Particle swarm optimization

来源：评论

学校读者我要写书评

暂无评论

Evaluation method of airport pavement partition coverage detection based on human-robot interaction

Evaluation method of airport pavement partition coverage det...

引用

Chinese Control and Decision Conference, CCDC

作者： Shilei Cheng Andong Shi Zhengxi Hu Kunxv Zhao Lei Sun Jingtai Liu Institute of Robotics and Automatic Information System Nankai University Tianjin Tianjin Key Laboratory of Intelligent Robotics Tianjin

ISBN: (纸本)9781665478977

The coverage detection of airport pavement is the premise to ensure civil aviation safety. Due to the hard requirements of ultra-high security of the airport itself, the application of robots to unmanned coverage detection in the airport is not accomplished overnight, but a process of step-by-step and continuous technical iteration. Therefore, for different application stages of the robot in the airport, an appropriate method should be given to guide airport maintenance engineers to flexibly adjust the partition coverage operation mode. In this paper, we design an evaluation index named Partition Quality Function (PQF) to evaluate the quality of partition coverage. In PQF, the degree suitable for human-robot interaction, the efficiency of coverage planning and the coverage failure area are taken into account. And the proportion of factors can be adjusted through the weight coefficient. Then, in order to facilitate engineering application, we design six kinds of rectangular partition templates and use genetic algorithm to optimize PQF. Finally, through simulation research, we discover PQF partition law under different weight coefficients, and gave the PQF setting of the robot in different application stages of the airport. It is an important step to promote the actual landing of robot coverage detection technology in the airport.

关键词： Atmospheric modeling Human-robot interaction Maintenance engineering Airports Safety Planning Security

来源：评论

学校读者我要写书评

暂无评论

Study of the shear strength behavior in flip chip under thermo-mechano-electrical coupling and different solder height

Study of the shear strength behavior in flip chip under ther...

引用

International Conference on (ICEPT) Electronic Packaging Technology

作者： Bin Zhou Jianxin Qiao Yunkang Su Dapeng Liu Wentao Huang Longqiu Li State Key Laboratory of Robotics and System Key Laboratory for Microsystems and Microstructure Manufacturing Harbin Institute of Technology Harbin China School of Mechanical and Electrical Engineering Harbin Institute of Technology Harbin China

Complicated service environment such as thermo-mechano-electrical coupling field and limited solder height have brought great challenges to the reliability of flip chips packaging and hindered the chip integration process. In this work, mechanical properties of 3.0Ag-0.5Cu(SAC305) solder under different service environment and height were investigated at macroscopic and microcosmic level. The flip chip shear test is realized on the universal test machine use the proposed fixture. The shear strength shows declining trend with the increment of coupling fields and solder heights. The underlying mechanism of strength loss is thermal softening effect and reduction rate could reach 42.77% at most. The Scanning Electron Microscopy (SEM) is used to analysis the fracture morphology and intermetallic compound (IMC) layer. The shear strength was negatively correlated with the thickness of IMC layer, and was positively correlated with fractal dimension of fracture morphology. The result shows that the thickness of IMC layer is the highest while the field is thermo-mechano-electrical and the solid height is 398μm. Meanwhile, the fractal dimension of fracture morphology is the lowest.

关键词：

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：