检索结果-内蒙古大学图书馆

EM-Gaze:eye context correlation and metric learning for gaze estimation

Visual Computing for Industry,Biomedicine,and Art 2023年第1期6卷 97-108页

作者： Jinchao Zhou Guoan Li Feng Shi Xiaoyan Guo Pengfei Wan Miao Wang State Key Laboratory of Virtual Reality Technology and Systems Beihang UniversityBeijing 100191China Kuaishou Technology Beijing 100085China

In recent years,deep learning techniques have been used to estimate gaze-a significant task in computer vision and human-computer *** studies have made significant achievements in predicting 2D or 3D gazes from monocular face *** study presents a deep neural network for 2D gaze estimation on mobile *** achieves state-of-the-art 2D gaze point regression error,while significantly improving gaze classification error on quadrant divisions of the *** this end,an efficient attention-based module that correlates and fuses the left and right eye contextual features is first proposed to improve gaze point regression ***,through a unified perspective for gaze estimation,metric learning for gaze classification on quadrant divisions is incorporated as additional ***,both gaze point regression and quadrant classification perfor-mances are *** experiments demonstrate that the proposed method outperforms existing gaze-estima-tion methods on the GazeCapture and MPIIFaceGaze datasets.

关键词： Computer vision Gaze estimation Metric learning Attention Multi-task learning

来源：评论

学校读者我要写书评

暂无评论

An Energy-Efficient Data Collection Scheme Using Denoising Autoencoder in Wireless Sensor Networks

引用

Tsinghua Science and technology 2019年第1期24卷 86-96页

作者： Guorui Li Sancheng Peng Cong Wang Jianwei Niu Ying Yuan School of Computer Science and Engineering Northeastern University School of Information Science and Technology Guangdong University of Foreign Studies State Key Laboratory of Virtual Reality Technology and Systems Beihang University

As one of the key operations in Wireless Sensor Networks(WSNs), the energy-efficient data collection schemes have been actively explored in the literature. However, the transform basis for sparsifing the sensed data is usually chosen empirically, and the transformed results are not always the sparsest. In this paper, we propose a Data Collection scheme based on Denoising Autoencoder(DCDA) to solve the above problem. In the data training phase, a Denoising AutoEncoder(DAE) is trained to compute the data measurement matrix and the data reconstruction matrix using the historical sensed data. Then, in the data collection phase, the sensed data of whole network are collected along a data collection tree. The data measurement matrix is utilized to compress the sensed data in each sensor node, and the data reconstruction matrix is utilized to reconstruct the original data in the ***, the data communication performance and data reconstruction performance of the proposed scheme are evaluated and compared with those of existing schemes using real-world sensed data. The experimental results show that compared to its counterparts, the proposed scheme results in a higher data compression rate, lower energy consumption, more accurate data reconstruction, and faster data reconstruction speed.

关键词： wireless sensor networks data collection neural networks autoencoder data reconstruction

来源：评论

学校读者我要写书评

暂无评论

Methods of projecting 2D wrinkle curves onto a 3D model

Methods of projecting 2D wrinkle curves onto a 3D model

引用

International Conference on Machine Tool technology and Mechatronics Engineering, ICMTTME 2014

作者： Lin, Xiao Dan He, Bing Zhang, Xiu Lin State Key Lab of Virtual Reality Technology and Systems Beihang University Beijing China

ISBN: (纸本)9783038352464

Wrinkles make cloth simulation results more realistic. However, generating wrinkles with physically based methods usually requires a computationally expensive simulation, while geometric methods such as deforming the 3D mesh based on predefined curves need time-consuming manual design. In this paper, we extract 2D wrinkle curves from pictures of clothing and project the curves onto a 3D clothing model. Two schemes for projecting 2D curves to the 3D model are presented in the paper, including the brute force method and the acceleration method. The brute force method has excellent projection results and the acceleration method enhances the efficiency of projection process significantly. A quad tree for the 3D model is built up in the acceleration method to improve the efficiency of projection process. © (2014) Trans Tech Publications, Switzerland.

关键词： Efficiency

来源：评论

学校读者我要写书评

暂无评论

A portable in-vivo device of friction force and torque measurement for vascular surgery

A portable in-vivo device of friction force and torque measu...

引用

2012 10th IEEE International Conference on virtual Environments, Human-Computer Interfaces, and Measurement systems, VECIMS 2012

作者： Wang, Dangxiao Yang, Cailing Shi, Jiangxia Zhang, Yuru State Key Lab of Virtual Reality Technology and Systems Beihang University Beijing China

ISBN: (纸本)9781457717574

In-vivo measurement of force signals between surgical tool and human tissues is an important research topic in high-fidelity surgical simulations. In this paper, we introduce a portable device for real-time in-vivo measurement of friction force and torque during vascular surgery operations. To measure multi-region contacts between the guide wire and the vascular wall, two portable sensor modules are mounted on surgeon's fingertip of thumb and index finger. Based on relative motion between an inner shaft and an outer socket, the sensor module can detect in real-time one-dimensional pushing or retracting resistance force and one-dimensional twisting resistance torque between the guide wire and the vascular wall. The sensor module is based on a modular design of three off-the-shelf one-dimensional force sensors, which provides a flexible solution to realize compact design. Preliminary experiments indicate the sensor modules can provide a portable solution for real-time measuring the force and torque signals during vascular surgery operations. © 2012 IEEE.

关键词： Friction

来源：评论

学校读者我要写书评

暂无评论

Modeling and editing isotropic BRDF

Modeling and editing isotropic BRDF

引用

International Conference on Modeling, Simulation and Visualization Methods (WMSVM)

作者： Hu, Yong Qi, Yue Bai, Gang State Key Lab of Virtual Reality Technology and Systems Beihang University Beijing China

ISBN: (纸本)9780769540467

Reflectance properties of real-world opaque materials can be represented by bidirectional reflectance distribution function (BRDF). Non-parametric BRDF becomes the main aspect nowadays because of its verisimilitude and generality. This paper introduces our BRDF capture equipment and a new compact non-parametric model based on curves for isotropic BRDF. 3D isotropic BRDF data are captured and assembled into a 2D matrix, then decomposed to several 1D curves through twice matrix decomposition, each curve has its own feature, such as specular shape curve, diffuse color curve, etc. The data are compressed effectively, and then reconstructed by these 1D curves during real-time rendering. More important, artist can edit the captured reflectance by adjusting one or more meaningful 1D curve intuitively. Experimental results show that, our method can capture the reflectance data, model and edit BRDF effectively. © 2010 IEEE.

关键词： Reflection

来源：评论

学校读者我要写书评

暂无评论

Interactive texture design and synthesis from mesh sketches

引用

Frontiers of Computer Science 2014年第2期8卷 330-341页

作者： Lili WANG Qinglin QI Yi CHEN Wei KE Aimin HAO State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing 100191 China Macao Polytechnic Institute Macao China

Geometry mesh introduces user control into tex- ture synthesis and editing, and brings more variations in the synthesized results. But still two problems related remain in need of better solutions. One problem is generating the meshes with desired size and pattern efficiently from easier user inputs. The other problem is improving the quality of synthesized results with mesh information. We present a new two-step texture design and synthesis method that addresses these two problems. Besides example texture, a small piece of mesh sketch drawn by hand or detected from example texture is input to our algorithm. And then a mesh synthesis method of geometry space is provided to avoid optimizations cell by cell. Distance and orientation features are introduced to im- prove the quality of mesh rasterization. Results show that with our method, users can design and synthesize textures from mesh sketches easily and interactively.

关键词： texture synthesis texture design feature-basedmethod structural texture

来源：评论

学校读者我要写书评

暂无评论

3D scene graph prediction from point clouds

引用

virtual reality & Intelligent Hardware 2022年第1期4卷 76-88页

作者： Fanfan WU Feihu YAN Weimin SHI Zhong ZHOU State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and EngineeringBeihang UniversityBeijing 100191China

Background In this study,we propose a novel 3D scene graph prediction approach for scene understanding from point *** It can automatically organize the entities of a scene in a graph,where objects are nodes and their relationships are modeled as *** specifically,we employ the DGCNN to capture the features of objects and their relationships in the scene.A Graph Attention Network(GAT)is introduced to exploit latent features obtained from the initial estimation to further refine the object arrangement in the graph structure.A one loss function modified from cross entropy with a variable weight is proposed to solve the multi-category problem in the prediction of object and *** Experiments reveal that the proposed approach performs favorably against the state-of-the-art methods in terms of predicate classification and relationship prediction and achieves comparable performance on object classification *** The 3D scene graph prediction approach can form an abstract description of the scene space from point clouds.

关键词： Scene understanding 3D scene graph Point cloud DGCNN GAT

来源：评论

学校读者我要写书评

暂无评论

Unmanned aerial systems coordinate target allocation based on wolf behaviors

引用

Science China(Information Sciences) 2019年第1期62卷 205-207页

作者： Haibin DUAN Qing YANG Yimin DENG Pei LI Huaxin QIU Tianjie ZHANG Daifeng ZHANG Mengzhen HUO Yankai SHEN Bio-inspired Autonomous Flight Systems (BAFS) Research Group State Key Laboratory of Virtual Reality Technology and SystemsSchool of Automation Science and Electrical EngineeringBeihang University (BUAA)

Unmanned aerial systems (UASs), especially the cluster of UAS, are highly focused and widely used in various domains. The application of small and inexpensive UAS with a large-scale group to perform complex tasks has ... 详细信息

关键词： UAS Unmanned aerial systems coordinate target allocation based on wolf behaviors

来源：评论

学校读者我要写书评

暂无评论

A survey on image and video stitching

引用

virtual reality & Intelligent Hardware 2019年第1期1卷 55-83页

作者： Wei LYU Zhong ZHOU Lang CHEN Yi ZHOU State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and EngineeringBeihang UniversityBeijing 100191China

Image/video stitching is a technology for solving the field of view(FOV)limitation of images/*** stitches multiple overlapping images/videos to generate a wide-FOV image/video,and has been used in various fields such as sports broadcasting,video surveillance,street view,and *** survey reviews image/video stitching algorithms,with a particular focus on those developed in recent *** stitching first calculates the corresponding relationships between multiple overlapping images,deforms and aligns the matched images,and then blends the aligned images to generate a wide-FOV image.A seamless method is always adopted to eliminate such potential flaws as ghosting and blurring caused by parallax or objects moving across the overlapping *** stitching is the further extension of image *** usually stitches selected frames of original videos to generate a stitching template by performing image stitching algorithms,and the subsequent frames can then be stitched according to the *** stitching is more complicated with moving objects or violent camera movement,because these factors introduce jitter,shakiness,ghosting,and *** detection technique is usually combined into stitching to eliminate ghosting and blurring,while video stabilization algorithms are adopted to solve the jitter and *** paper further discusses panoramic stitching as a special-extension of image/video *** stitching is currently the most widely used application in *** survey reviews the latest image/video stitching methods,and introduces the fundamental principles/advantages/weaknesses of image/video stitching ***/video stitching faces long-term challenges such as wide baseline,large parallax,and low-texture problem in the overlapping *** technologies may present new opportunities to address these issues,such as deep learning-based semantic correspondence,and 3D image ***,

关键词： Image stitching Video stitching Panoramic stitching Registration Alignment Mesh optimization Deep learning 3D stitching

来源：评论

学校读者我要写书评

暂无评论

Multi-view stereo reconstruction for internet photos

Multi-view stereo reconstruction for internet photos

引用

2011 International Conference on virtual reality and Visualization, ICVRV 2011

作者： Yu, Sijia Qi, Yue Shen, Xukun State Key Lab. of Virtual Reality Technology and Systems Bei Hang University China

ISBN: (纸本)9780769546025

This paper develops a multi-view stereo approach to reconstruct the shape of a 3D object from a set of Internet photos. The stereo matching technique adopts region growing approach, starting from a set of sparse 3D points reconstructed from structure-from-motion (Sfm), then propagates to the neighbouring areas by the best-first strategy, and produces dense 3D points. View selection and filter algorithms are proposed considering the characteristics of Internet images. Specifically, for each 3D point, we choose a reference image at first which determines the right subset images from unordered sets for optimization. Two filter algorithms, namely Sfm points filter and quality filter which is based on the assumption that depth changes smoothly, are designed to eliminate low-quality reconstructions. We demonstrate our algorithms with several datasets which show that they perform robustly. © 2011 IEEE.

关键词： Bandpass filters

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：