检索结果-南通市图书馆

Multi-scale object detection by top-down and bottom-up feature pyramid network

在线全文

学校读者我要写书评

暂无评论

Journal of Systems Engineering and Electronics 2019年第1期30卷 1-12页

作者： ZHAO Baojun ZHAO Boya TANG Linbo WANG Wenzheng WU Chen School of Information and Electronics Beijing Institute of Technology Beijing Key Laboratory of Embedded Real-time Information Processing Technology Beijing Institute of Technology

While moving ahead with the object detection technology, especially deep neural networks, many related tasks, such as medical application and industrial automation, have achieved great success. However, the detection of objects with multiple aspect ratios and scales is still a key problem. This paper proposes a top-down and bottom-up feature pyramid network(TDBU-FPN),which combines multi-scale feature representation and anchor generation at multiple aspect ratios. First, in order to build the multi-scale feature map, this paper puts a number of fully convolutional layers after the backbone. Second, to link neighboring feature maps, top-down and bottom-up flows are adopted to introduce context information via top-down flow and supplement suboriginal information via bottom-up flow. The top-down flow refers to the deconvolution procedure, and the bottom-up flow refers to the pooling procedure. Third, the problem of adapting different object aspect ratios is tackled via many anchor shapes with different aspect ratios on each multi-scale feature map. The proposed method is evaluated on the pattern analysis, statistical modeling and computational learning visual object classes(PASCAL VOC)dataset and reaches an accuracy of 79%, which exhibits a 1.8% improvement with a detection speed of 23 fps.

关键词： convolutional neural network (CNN) feature pyramid network (FPN) object detection deconvolution.

Preparation of pyramid-SiNWs binary structure with Ag nanoparticles-assisted chemical etching

在线全文

学校读者我要写书评

暂无评论

Rare Metals 2019年第4期38卷 312-315页

作者： Zi-Long Zhang Bo Wang Yu Chen Yun-Hui Tang Xue-Mei Song Qing-Liu Li Hui Yan Department of Materials Science and Engineering Beijing University of Technology

It is an important way to improve the efficiency of solar cells by using the special microstructures of *** this work,a pyramid-silicon nanowires (pyramidSiNWs) binary structure was prepared on the silicon surfaces with the metal-assisted chemical etching (MACE) *** electron microscope (SEM) was used to observe the micromorphology of the pyramid-SiNWs binary *** formation mechanism of the binary structure was *** role of Ag nanoparticles in MACE is considered to be the template and the *** optical reflectivity of the silicon surfaces was studied with ultraviolet-visible (UV-Vis) *** with the fiat silicon surface and the simple pyramidal structure,the silicon surfaces with the pyramid-SiNWs binary structure achieve a much lower reflectance in a wide range of *** effect of etching time as a parameter on the reflectivity was also discussed.

关键词： Ag nanoparticle pyramid structure Silicon nanowires Binary structure Reflectance

Gender-Specific Multi-Task Micro-Expression Recognition Using pyramid CGBP-TOP Feature

在线全文

学校读者我要写书评

暂无评论

Computer Modeling in Engineering & Sciences 2019年第3期118卷 547-559页

作者： Chunlong Hu Jianjun Chen Xin Zuo Haitao Zou Xing Deng Yucheng Shu School of Computer Science and Engineering Jiangsu University of Science and TechnologyZhenjiang212003Chinal School of Computer Science and Technology Chongqing University of Posts and TelecommunicationsChongqing400065China

Micro-expression recognition has attracted growing research interests in the field of compute ***,micro-expression usually lasts a few seconds,thus it is difficult to *** paper presents a new framework to recognize micro-expression using pyramid histogram of Centralized Gabor Binary Pattern from Three Orthogonal Panels(CGBP-TOP)which is an extension of Local Gabor Binary Pattern from Three Orthogonal Panels ***-TOP performs spatial and temporal analysis to capture the local facial characteristics of micro-expression image *** order to keep more local information of the face,CGBP-TOP is extracted based on pyramid subregions of the micro-expression video *** combination of CGBP-TOP and spatial pyramid can represent well and truly the facial movements of the micro-expression image ***,the dimension of our pyramid CGBP-TOP tends to be very high,which may lead to high data redundancy *** addition,it is clear that people of different genders usually have different ways of ***,in this paper,in order to select the relevant features of micro-expression,the gender-specific sparse multi-task learning method with adaptive regularization term is adopted to learn a compact subset of pyramid CGBP-TOP feature for micro-expression classification of different ***,extensive experiments on widely used CASME II and SMIC databases demonstrate that our method can efficiently extract micro-expression motion features in the micro-expression video ***,our proposed approach achieves comparable results with the state-of-the-art methods.

关键词： Micro-expression recognition feature extraction spatial pyramid multi-task learning regularization

Macro-pyramid in GaN Film

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics Letters 2001年第2期18卷 260-262页

作者： ZHOU Jing YANG Zhi-Jian XU Shi-Fa ZHU Xing ZHANG Guo-Yi Mesoscopic Physics National Laboratory and Department of Physics Peking UniversityBeijing 100871

A thin film of GaN with the thickness of 1.0μm was grown onα-Al2O3 substrate by metal organic chemical vapour disposition and then a thick GaN film with thickness of 12μm was grown in the halide vapour phase epitaxy *** macro-pyramids appeared on the surface of the *** macro-pyramids made the surface of the GaN film rough,which was harmful to the devices made by GaN *** defects changed the distribution of carrier concentration and affected the optical properties of *** step height of the pyramids was about 30-40 nm measured by atomic force microscopy.A simple model was proposed to explain the macro-pyramid phenomenon compared with the growth *** growth of the macro-pyramid was relative to the physical conditions in the reaction *** increasing growth temperature and low pressure may reduce the pyramid size.

关键词： GaN 2μm pyramid

BFP Net: Balanced Feature pyramid Network for Small Apple Detection in Complex Orchard Environment

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Plant Phenomics 2022年第1期4卷 75-93页

作者： Meili Sun Liancheng Xu Xiude Chen Ze Ji Yuanjie Zheng Weikuan Jia School of Information Science and Engineering Shandong Normal UniversityJinanChina Key Laboratory of Facility Agriculture Measurement and Control Technology and Equipment of Machinery Industry Zhenjiang 212013China National Engineering Research Center for Apple Shandong Agriculture UniversityTaian 271018China School of Engineering Cardiff UniversityCardiff CF243AAUK

Despite of significant achievements made in the detection of target fruits,small fruit detection remains a great challenge,especially for immature small green fruits with a few *** closeness of color between the fruit skin and the background greatly increases the difficulty of locating small target fruits in the natural orchard *** this paper,we propose a balanced feature pyramid network(BFP Net)for small apple detection.

关键词： Net pyramid locating

A two‐branch network with pyramid‐based local and spatial attention global feature learning for vehicle re‐identification

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

CAAI Transactions on Intelligence Technology 2021年第1期6卷 46-54页

作者： Jucheng Yang Di Xing Zhiqiang Hu Tong Yao College of Artificial Intelligence Tianjin University of Science and TechnologyTianjinPeople’s Republic of China

In recent years,vehicle re‐identification has attracted more and more *** to learn the discriminative information from multi‐view vehicle images becomes one of the challenging problems in vehicle re‐identification *** example,when the viewpoint of the image changes,the features extracted from one image may be lost in another image.A two‐branch network with pyramid‐based local and spatial attention global feature learning(PSA)is proposed for vehicle re‐identification to solve this ***,one branch learns local features at different scales by building pyramid from coarse to fine and the other branch learns attentive global features by using spatial attention ***,pooling operation by using global maximum pooling(GMP)for local features and global average pooling(GAP)for global feature is ***,local feature vectors and global feature vector extracted from the last pooling layer,respectively,are employed for identity re‐*** experimental results demonstrate that the proposed method achieves state‐of‐the‐art results on the VeRi‐776 dataset and VehicleID dataset.

关键词： branch pyramid learning

Fabrication of dodecagonal pyramid on nitrogen face GaN and its effect on the light extraction

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Science China(Technological Sciences) 2010年第3期53卷 769-771页

作者： QI ShengLi,CHEN ZhiZhong,SUN YongJian,FANG Hao,TAO YueBin,SANG LiWen,TIAN PengFei,DENG JunJing,ZHAO LuBing,YU TongJun,QIN ZhiXin&ZHANG GuoYi State Key Laboratory for Artificial Microstructure and Mesoscopic Physics,School of Physics,Peking University,Beijing 100871,China 1. State Key Laboratory for Artificial Microstructure and Mesoscopic Physics School of Physics Peking University Beijing 100871 China

Wet etching has been widely used in defect evaluation for Ga-face GaN and surface roughness for N-face GaN dodecagonal pyramids has been fabricated on laser-lift-off N-face GaN by hot phosphor acid *** dodecagonal pyramid shows twelve facets including six{20-2-3}and six{22-4-5}*** cross-sectional TEM image,it is shown that the pyramid corresponds to the top of the edge *** with hexagonal pyramid-surface light emitting diodes(LEDs)etched by commonly used photoelectrochemical(PEC)process in KOH aqueous,the dodecagonal pyramid-surface LEDs show improved light extraction efficiency because of more facets,which effectively reduces the total internal reflection.

关键词： GaN N face wet etching H3PO4 dodecagonal pyramid light extraction

Distributed GIS oriented generalized image pyramid and its practice

在线全文

学校读者我要写书评

暂无评论

Science China(Technological Sciences) 2010年第S1期53卷 99-104页

作者： XIA YuBin,PANG PeiYu & WENG JingNong School of Computer Science and Engineering,Beihang University,Beijing 100191,China 1. School of Computer Science and Engineering Beihang University Beijing 100191 China

Gaussian pyramid framework (GPF) and Laplacian pyramid framework (LPF) are the two main frameworks used for transmission and structure of spatial image and geometric *** is the main approach for progressive transmission of spatial image *** paper discusses the issues in structure,transmission and storage efficiency of spatial data in the current spatially distributed visualization *** in the paper a novel approach is introduced to structure and store spatial data based on the *** new approach builds (spatial) index for transformed frequency data within different ranges based on splitting high-frequency and low-frequency data by biorthogonal *** approach decreases the storage consumption and network data traffic for transmission while enables the progressive transmission of spatial image.

关键词： redundancy spatial index progressive transmission generalized pyramid structure Laplacian pyramid framework

A 2D quadrangular pyramid photoelectric autocollimator with extended angle measurement range

在线全文

学校读者我要写书评

暂无评论

Optoelectronics Letters 2021年第8期17卷 468-474页

作者： Konyakhin Igor LI Renpu ZHOU Min Dang Dinh Duan Nikitin Mikhail HUANG Guifu YANG Jiawen TAN Xin Faculty of Applied Optic 1TMO University49 Kronverkskiy ProspectSaint Petersburg 197101Russia Chongqing Engineering Research Center of Intelligent Sensing Technology and Microsystem Chongqing University of Post and TelecommunicationsChongqing 400065China School of Communication and Information Engineering Chongqing University of Posts and TelecommunicationsChongqing 400065China

A photoelectric autocollimator with high accuracy and extended measurement range based on the quadrangular pyramid is proposed,and the corresponding algorithms are also deduced.A new image processing algorithm has been proposed to improve the accuracy,and the corresponding errors are also estimated,the error does not exceed half a pixel when the distance between the marks more than two *** experimental results have verified that the measurement range of the proposed two-dimensional(2D)quadrangular pyramid photoelectric autocollimator can be increased 2 times than that of the flat mirror photoelectric autocollimator from 10′to 15′.The accuracy is better than 1″when the deflection is less than 15′.

关键词： tocol pyramid photoelectric