检索结果-南通市图书馆

Bidirectional Optimization Coupled Ligh tweight Net works for Efficient and Robust Multi-Person 2D pose Estimation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Computer Science & Technology 2019年第3期34卷 522-536页

作者： Shnai Li Zheng Fang Wen-Feng Song Ai-Min Hao Hong Qin State Key Laboratory of Virtual Reality Technology and Systems Beihang UniversityBeijing 100191China Beihang University Qingdao Research Institute Qingdao 266000China Department of Computer Science Stony Brook UniversityStony Brook 11790U.S.A.

For multi-person 2D pose estimation,current deep learning baised methods have exhibited impressive performance,but the trade-offs among efficiency,robustness,and accuracy in the existing approaches remain *** principle,bottom-up methods are superior to top-down methods in efficiency,but they perform worse in *** make full use of their respective advantages,in this paper we design a novel bidirectional optimization coupled lightweight network(BOCLN)architecture for efficient,robust,and general-purpose multi-person 2D(2-dimensional)pose estimation from natural *** the BOCLN framework,the bottom-up network focuses oil global features,while the top-down net work places emphasis on det ailed *** entire framework shares global features along the bottom-up data stream,while the top-down data stream aims to accelerate the accurate pose *** particular,to exploit the priors of human joints'relationship,we propose a probability limb heat map to represent the spatial context of the joints and guide the overall pose skeleton prediction,so that each person's pose estimation in cluttered scenes(involving crowd)could be as accurate and robust as ***,benefiting from the novel BOCLN architecture,the tinie-consuming refinement procedure could be much simplified to an efficient lightweight *** experiments and evaluations on public benchmarks have confirmed that our new method is more efficient and robust,yet still attain competitive accuracy performance compared with the state-of-the-art *** BOCLN shows even greater promise in online applications.

关键词： bidirectional optimization computer vision deep learning probability limb heat map 2D multi-person pose estimation

Real-Time Avatar pose Transfer and Motion Generation Using Locally Encoded Laplacian Offsets

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Computer Science & Technology 2019年第2期34卷 256-271页

作者： Masoud Zadghorban Lifkooee Celong Liu Yongqing Liang Yimin Zhu Xin Li Department of Electrical and Computer Engineering Louisiana State University Baton Rouge 70803 U.S.A. Department of Construction Management Louisiana State University Baton Rouge 70803 U.S.A.

We propose a human avatar representation scheme based on intrinsic coordinates,which are invariant to isometry and insensitive to human pose changes,and an efficient pose transfer algorithm that can utilize this representation to reconstruct a human body geometry following a given *** a pose transfer algorithm can be used to control the movement of an avatar model in virtual reality environments following a user's motion in real *** proposed algorithm consists of three main ***,we recognize the user's pose and select a template model from the database who has a similar pose;then,the intrinsic Laplacian offsets encoded in local coordinates are used to reconstruct the human body geometry following the template pose;finally,the morphing between the two poses is generated using a linear *** perform experiments to evaluate the accuracy and efficiency of our *** believe our proposed system is a promising human modeling tool that can be used in general virtual reality applications.

关键词： human body pose transfer local intrinsic coordinates avatar control in virtual reality

Deep Learning Based 2D Human pose Estimation:A Survey

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Tsinghua Science and Technology 2019年第6期24卷 663-676页

作者： Qi Dang Jianqin Yin Bin Wang Wenqing Zheng School Beijing University of Posts and TelecommunicationsBeijing 100876China State Key Lab.of Intelligent Technology and Systems Tsinghua UniversityBeijing 100084China School of Information and Communication Engineering Beijing University of Posts and Telecommunications Beijing 100876China

Human pose estimation has received significant attention recently due to its various applications in the real world. As the performance of the state-of-the-art human pose estimation methods can be improved by deep learning, this paper presents a comprehensive survey of deep learning based human pose estimation methods and analyzes the methodologies employed. We summarize and discuss recent works with a methodologybased taxonomy. Single-person and multi-person pipelines are first reviewed separately. Then, the deep learning techniques applied in these pipelines are compared and analyzed. The datasets and metrics used in this task are also discussed and compared. The aim of this survey is to make every step in the estimation pipelines interpretable and to provide readers a readily comprehensible explanation. Moreover, the unsolved problems and challenges for future research are discussed.

关键词： human pose estimation deep learning computer vision

Precise monocular vision-based pose measurement system for lunar surface sampling manipulator

在线全文

学校读者我要写书评

暂无评论

Science China(Technological Sciences) 2019年第10期62卷 1783-1794页

作者： WANG Gang SHI ZhongChen SHANG Yang SUN XiaoLiang ZHANG WenLong YU QiFeng College of Aerospace Science and Engineering National University of Defense TechnologyChangsha410073China Hunan Provincial Key Laboratory of Image Measurement and Vision Navigation Changsha410073China

Space manipulator has been playing an increasingly important role in space exploration due to its flexibility and versatility. This paper is to design a vision-based pose measurement system for a four-degree-of-freedom(4-DOF) lunar surface sampling manipulator relying on a monitoring camera and several fiducial markers. The system first employs double plateaus histogram equalization for the markers to improve the robustness to varying noise and illumination. The markers are then accurately extracted in sub-pixel based on template matching and curved surface fitting. Finally, given the camera parameters and 3D reference points, the pose of the manipulator end-effector is solved from the 3D-to-2D point correspondences by combining a plane-based pose estimation method with rigid-body transformation. Experiment results show that the system achieves highprecision positioning and orientation performance. The measurement error is within 3 mm in position, and 0.2° in orientation,meeting the requirements for space manipulator operations.

关键词： vision-based pose measurement pose estimation lunar surface sampling sub-pixel corner detection Chang’e 5

pose Estimation of Space Targets Based on Model Matching for Large-Aperture Ground-Based Telescopes

在线全文

学校读者我要写书评

暂无评论

Computer Modeling in Engineering & Sciences 2018年第11期117卷 271-286页

作者： Zhengwei Li Jianli Wang Tao Chen Bin Wang Yuanhao Wu Changchun Institute of Optics Fine Mechanics and PhysicsChinese Academy of SciencesChangchun13033China University of Chinese Academy of Sciences Beijing100049China.

With the development of adaptive optics and post restore processing techniques,large aperture ground-based telescopes can obtain high-resolution images(HRIs)of *** pose of the space target can be estimated from HRIs by several *** the target features obtained from the image are unstable,it is difficult to use existing methods for pose *** this paper a method based on real-time target model matching to estimate the pose of space targets is ***,the physicallyconstrained iterative deconvolution algorithm is used to obtain HRIs of the space ***,according to the 3D model,the ephemeris data,the observation time of the target,and the optical parameters of the telescope,the simulated observation image of the target in orbit is rendered by a scene simulation ***,the target model searches through yaw,pitch,and roll until the correlation between the simulated observation image and the actual observation image shows an optimal *** simulation results show that the proposed pose estimation method can converge to the local optimal value with an estimation error of about 1.6349°.

关键词： Ground-based telescope pose estimation correlation matching space target image restoration

Hardware-in-loop simulation on hydrostatic thrust bearing worktable pose

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Central South University 2008年第S2期15卷 250-256页

作者：韩桂华邵俊鹏秦柏董玉红 Mechanical Power Engineering College Harbin University of Science and Technology

A controllable hydrostatic thrust bearing was presented to improve rigidity. The bearing worktable poses were controlled by coupling oilfilm thickness of four controllable chambers. The chamber flow can be regulated by electro hydraulic servo valve-control variable pump according to the surface roughness, load, cutting force, and thermal effects of worktable. The mathematical models of the controllable chamber flow, servo variable mechanism and controller were built. The pose control model was established, which contained the kinematics positive and negative solution and control strategy of feedforward and hydraulic cylinder position feedback. Hardware-in-loop simulation experiment was carried out on the electro hydraulic servo test bench by means of the non-linear relation of film thickness and hydraulic cylinder displacement. Hardware-in-loop simulation experiment results show that the controllable bearings exhibit high oilfilm rigidity, the rising time is 0.24 s and the maximum overshoot is 2.23%, and can be applied in high precision heavy machine tool.

关键词： hydrostatic thrust bearing hardware-in-loop simulation worktable pose controllable chamber

Multi-Residual Module Stacked Hourglass Networks for Human pose Estimation

在线全文

学校读者我要写书评

暂无评论

Journal of Beijing Institute of Technology 2020年第1期29卷 110-119页

作者： Wenxia Bao Yaping Yang Dong Liang Ming Zhu Key Laboratory Intelligent Computing and Signal Processing of the Ministry of Education Anhui UniversityHefei 230039China

A multi-residual module stacked hourglass network(MRSH)was proposed to improve the accuracy and robustness of human body pose *** network uses multiple hourglass sub-networks and three new residual *** the hourglass sub-network,the large receptive field residual module(LRFRM)and the multi-scale residual module(MSRM)are first used to learn the spatial relationship between features and body parts at various *** the improved residual module(IRM)is used when the resolution is *** final network uses four stacked hourglass sub-networks,with intermediate supervision at the end of each hourglass,repeating high-low(from high resolution to low resolution)and low-high(from low resolution to high resolution)*** network was tested on the public datasets of Leeds sports poses(LSP)and MPII human *** experimental results show that the proposed network has better performance in human pose estimation.

关键词： human pose estimation residual learning image feature hourglass network

DETERMINING pose PARAMETERS WITHOUT PRIMITIVE-TO-PRIMITIVE CORRESPONDENCE

在线全文

学校读者我要写书评

暂无评论

Journal of Electronics(China) 1995年第2期12卷 113-123页

作者：曹钧何振亚 Department of Computer Science & Technology Tsinghua University Beijing 100084 Department of Radio Engineering Southeast University Nanjing 210018

The problem of determining the pose of an object in 3-D space is essential in many computer vision applications. In this paper, a model-based approach for solving this problem is proposed. This approach does not require the knowledge of point-to-point correspondences between 3-D points on the model and 2-D points in the observed image. The spatial location of the object is iteratively estimated and updated from the values globally defined over the model image and the observed image.

关键词： pose determination Object localization Image processing Computer vision

Learning Hand Latent Features for Unsupervised 3D Hand pose Estimation

在线全文

学校读者我要写书评

暂无评论

Journal of Autonomous Intelligence 2019年第1期2卷 1-10页

作者： Jamal Banzi Isack Bulugu Zhongfu Ye School of Information Science and Technology University of Science and Technology of China230026China Sokoine University of Agriculture Morogoro3167Tanzania College of information and communication Technology University of Dare-es-salaamDar-es-Salaam33335Tanzania

Recent hand pose estimation methods require large numbers of annotated training data to extract the dynamic information from a hand ***,precise and dense annotation on the real data is difficult to come by and the amount of information passed to the training algorithm is significantly *** paper presents an approach to developing a hand pose estimation system which can accurately regress a 3D pose in an unsupervised *** whole process is performed in three ***,the hand is modelled by a novel latent tree dependency model (LTDM) which transforms internal joints location to an explicit ***,we perform predictive coding of image sequences of hand poses in order to capture latent features underlying a given image without supervision.A mapping is then performed between an image depth and a generated ***,the hand joints are regressed using convolutional neural networks to finally estimate the latent pose given some depth ***,an unsupervised error term which is a part of the recurrent architecture ensures smooth estimation of the final *** demonstrate the performance of the proposed system,a complete experiment was conducted on three challenging public datasets,ICVL,MSRA,and *** empirical results show the significant performance of our method which is comparable or better than the state-of-the-art approaches.

关键词： Hand pose Estimation Convolutional Neural Networks Recurrent Neural Networks Human-machine Interaction Predictive Coding Unsupervised Learning