检索结果-南通市图书馆

Depth-Guided Vision Transformer With Normalizing Flows for Monocular 3D Object Detection

在线全文

学校读者我要写书评

暂无评论

IEEE/CAA Journal of Automatica Sinica 2024年第3期11卷 673-689页

作者： Cong Pan Junran peng Zhaoxiang Zhang the Center for Research on Intelligent Perception and Computing(CRIPAC) National Laboratory of Pattern Recognition(NLPR)Institute of AutomationChinese Academy of Sciences(CASIA)Beijing 100190 the School of Future Technology University of Chinese Academy of Sciences(UCAS)Beijing 100049China the Huawei Inc. Beijing 100085China IEEE the Institute of Automation Chinese Academy of Sciences(CASIA)Beijing 100190 the University of Chinese Academy of Sciences(UCAS) Beijing 100049 the Centre for Artificial Intelligence and Robotics Hong Kong Institute of Science&Innovation Chinese Academy of Sciences(HKISI CAS)Hong Kong 999077China

Monocular 3D object detection is challenging due to the lack of accurate depth *** methods estimate the pixel-wise depth maps from off-the-shelf depth estimators and then use them as an additional input to augment the RGB ***-based methods attempt to convert estimated depth maps to pseudo-LiDAR and then use LiDAR-based object detectors or focus on the perspective of image and depth fusion ***,they demonstrate limited performance and efficiency as a result of depth inaccuracy and complex fusion mode with *** from these approaches,our proposed depth-guided vision transformer with a normalizing flows(NF-DVT)network uses normalizing flows to build priors in depth maps to achieve more accurate depth *** we develop a novel Swin-Transformer-based backbone with a fusion module to process RGB image patches and depth map patches with two separate branches and fuse them using cross-attention to exchange information with each ***,with the help of pixel-wise relative depth values in depth maps,we develop new relative position embeddings in the cross-attention mechanism to capture more accurate sequence ordering of input *** method is the first Swin-Transformer-based backbone architecture for monocular 3D object *** experimental results on the KITTI and the challenging Waymo Open datasets show the effectiveness of our proposed method and superior performance over previous counterparts.

关键词： Monocular 3D object detection normalizing flows Swin Transformer

Morphology of impact fragmentation distribution of single spherical and ellipsoidal particles in drop weight experiments

在线全文

学校读者我要写书评

暂无评论

Particuology 2024年第3期86卷 137-148页

作者： Qiuhua Miao peng Huang Wenchang Zhu School of Mechanical Engineering Southeast UniversityNanjing211189China

Particle shape is an important factor affecting the fragmentation distribution of the ore *** investigate the influence of particle shape on the morphological fragmentation distribution characteristics,the crushable ore particles are defined as prolate,oblate ellipsoid and spherical particles,which have different aspect ratios(AR)and sphericity(S).Based on the drop weight experiment,the influence of the net drop height on the macroscopic mechanical behavior and crushing distribution characteristics of the single spherical and ellipsoidal particles is *** results show that different peak-shifting characteristics exist during particle *** fragmentation distribution peak shifts left when the increased impact energy is eventually only enough to break medium-sized ***,it shifts right when impact energy is increased enough to break largest-sized ***,regardless of whether the net drop height changes,the maximum continuous fragmentation degree presents"M"-shaped characteristic with the increased *** with the ellipsoid particles,the single spherical particle is more difficult to be broken by impact,with wider equivalent particle fragmentation *** the increase of particle sphericity,the maximum continuous fragmentation degree of a single ellipsoid particle has an overall trend of initial increase and subsequent *** when particle sphericity is 0.9

关键词： Ellipsoid particles Fragmentation distribution Drop weight experiment Numerical simulation

Unsupervised construction of health indicator for rotating machinery via multi-criterion feature selection and attentive variational autoencoder

维普期刊数据库评论

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) 2024年第5期67卷 1524-1537页

作者： LI XinYu CHENG ChangMing peng ZhiKe State Key Laboratory of Mechanical System and Vibration Shanghai Jiao Tong UniversityShanghai200240China School of Mechanical Engineering Ningxia UniversityYinchuan750021China

Health indicator(HI)construction is a crucial task in degradation evaluation and facilitates the prognostic and health management(PHM)of rotating *** interference from artificial labeling,the HI construction approaches in an unsupervised manner have attracted substantial ***,current unsupervised methods generally struggle with two problems:(1)ignorance of both redundancy between features and global variability of features during the feature selection process;(2)inadequate utilization of information from different sampling *** tackle these problems,this work develops a novel unsupervised approach for HI construction that integrates multi-criterion feature selection and the Attentive Variational Autoencoder(Attentive VAE).Explicitly,a multi-criterion feature selection(Mc FS)algorithm together with an elaborately designed metric is proposed to determine a superior feature subset,considering the relevance,the redundancy,and the global variability of features ***,for the adequate utilization of the information from distinct sampling moments,a deep learning model named Attentive VAE is *** Attentive VAE is solely fed with the selected features in the health state for model training and the HI is derived through the reconstruction error to reveal the degradation degree of *** case studies based on genuine experimental datasets are involved to quantitatively evaluate the superiority of the developed approach,demonstrating its superiority over other unsupervised methods for characterizing degradation *** effectiveness of both the Mc FS algorithm and the Attentive VAE is verified by ablation experiments,respectively.

关键词： health indicator(HI) unsupervised learning multi-criterion feature selection global variability attention mechanism

Resource Allocation to Minimize the Makespan with Multi-Resource Operations

在线全文

学校读者我要写书评

暂无评论

Journal of Systems Science & Complexity 2024年第5期37卷 2054-2070页

作者： WENG Wuyan CHU Chengbin WU peng School of Economics and Management Fuzhou University Univ Gustave Eiffel ESIEE Paris&COSYS-GRETTIA

This paper investigates a new resource-allocation problem involving multi-resource operations, where completing an operation requires simultaneous use of multiple(renewable) resources,probably of different types. The goal of the study is to provide a solution method that minimizes the makespan. The authors formulate the problem into a novel mixed-integer linear program(MILP)model. To efficiently solve practical-sized instances, an exact Benders decomposition algorithm is developed. This algorithm divides the original problem into a master problem of allocating resources and a subproblem of calculating the makespan, and both are linked via Benders cuts. The convergence is sped up by improving the mathematical model and embedding the variable neighborhood search algorithm. Compared with CPLEX, a commonly used MILP solver, the computational results demonstrate that the proposed algorithm provides tighter upper and lower bounds in most instances. In particular,compared with CPLEX, the proposed method can on average improve the upper and lower bounds by 4.76% and 4.39%, respectively, in solving practical-sized instances.

关键词： Benders decomposition multi-resources operations resource allocation

Cross-Dimension Attentive Feature Fusion Network for Unsupervised Time-Series Anomaly Detection

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Computer Modeling in Engineering & Sciences 2024年第6期139卷 3011-3027页

作者： Rui Wang Yao Zhou Guangchun Luo peng Chen Dezhong peng School of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu611731China School of Computer and Software Engineering Xihua UniversityChengdu610039China School of Computer Science Sichuan UniversityChengdu610065China National Innovation Center for UHD Video Technology Chengdu610095China

Time series anomaly detection is crucial in various industrial applications to identify unusual behaviors within the time series *** to the challenges associated with annotating anomaly events,time series reconstruction has become a prevalent approach for unsupervised anomaly ***,effectively learning representations and achieving accurate detection results remain challenging due to the intricate temporal patterns and dependencies in real-world time *** this paper,we propose a cross-dimension attentive feature fusion network for time series anomaly detection,referred to as ***,a series and feature mixing block is introduced to learn representations in 1D ***,a fast Fourier transform is employed to convert the time series into 2D space,providing the capability for 2D feature ***,a cross-dimension attentive feature fusion mechanism is designed that adaptively integrates features across different dimensions for anomaly *** results on real-world time series datasets demonstrate that CAFFN performs better than other competing methods in time series anomaly detection.

关键词： Time series anomaly detection unsupervised feature learning feature fusion

Enhanced potassium storage of carbon nanofibers as binderfree anodes enabled by coupling ultra-small amorphous Sb_(2)O_(3),graphene modification and sulfur doping

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Rare Metals 2024年第1期43卷 51-64页

作者： Zhao Huang Yu-Kang Lou Lin peng Yuan peng Meng-Meng Wang Ming Zhang College of Railway Transportation Hunan University of TechnologyZhuzhou 412007China College of Semiconductors(College of Integrated Circuits) Changsha Semiconductor Technology and Application Innovation Research InstituteHunan UniversityChangsha 410082China Key Laboratory for Micro-Nano Optoelectronic Devices of Ministry of Education College of Semiconductors(College of Integrated Circuits)Hunan UniversityChangsha 410082China

Considering the intrinsic advantages of natural copiousness and cost-effectiveness of potassium resource,potassium-ion batteries(KIBs) are booming as prospective alternatives to lithium-ion batteries(LIBs) in large-scale energy storage scenarios. Nevertheless, lacking desirable electrodes for reversibly hosting the bulky K+hinders the widespread application of KIBs, and it needs to be urgently solved. Hereon, the porous S-doped Sb_(2)O_(3)-graphene-carbon(SAGC) nanofibers are manufactured through an adjustable and facile approach, which involves electrospinning, in situ etching and sulfuration. The synthesized SAGC is featured by the ultra-small amorphous Sb_(2)O_(3) homogeneously wrapped inside the carbon matrix, as well as the co-incorporation of graphene and sulfur. Tentatively,the SAGC nanofiber sheets are applied as binder-free anodes for KIBs, exhibiting a prominent cycling life(256.72 m Ah·g^(-1) over 150 cycles at 100 m A·g^(-1)) and rate·g^(-1) over 100 cycles at 1 A·g^(-1)). The positive synergy among all the active components accounts for the distinguished performances of the SAGC. By reinforcing the tolerability to the swelling stress, producing the valid electrochemical active sites, and promoting the charge transferring for reversible K+uptake, the SAGC finally renders the excellent cyclability, capacity, and rate capability. Moreover, the extrinsic electrochemical pseudocapacitance characteristics induced by the porous carbon substrate elevate the K-storage capacity of the SAGC as well. It is hoped that the conclusions drawn may offer new insights into a direction for the high-performance binderfree KIB anodes.

关键词： S-doped Sb_(2)O_(3)-graphene-carbon Nanofibers Binder-free Anodes Potassium-ion batteries

Stability assessment of tree ring growth of Pinus armandii Franch in response to climate change based on slope directions at the Lubanling in the Funiu Mountains,China

在线全文

学校读者我要写书评

暂无评论

Journal of Forestry Research 2024年第3期35卷 197-208页

作者： Jinkuan Li Jianfeng peng Xiaoxu Wei Meng peng Xuan Li Yameng Liu Jiaxin Li College of Geography and Environmental Science Henan UniversityKaifeng 475004People’s Republic of China Key Laboratory of Earth System Observation and Simulation of Henan Province Kaifeng 475004People’s Republic of China

Global warming will affect growth strategies and how trees will *** compare the response of tree radial growth to climate warming in different slope directions,samples of Pinus armandii Franch were collected and treering chronologies developed on northern and western slopes from the Lubanling in the Funiu *** analyses showed that two chronologies were mainly limited by temperatures in the previous June–August and the combination of temperatures and moisture in the current May–*** difference of the climate response to slopes was small but not *** growth of the LBL01 site on the northern slope was affected by the combined maximum and minimum temperatures,while that of the LBL02 site was affected by maximum *** regards to moisture,radial growth of the trees on the north slope was influenced by the relative humidity in the current May–July,while on the western slope,it was affected by the relative humidity in the previous June–August,the current May–July and the precipitation in the current May–*** the change in climate,the effects of the main limiting factors on growth on different slopes were visible to a certain extent,but the differences in response of trees on different slopes gradually decreased,which might be caused by factors such as different slope directions and the change in diurnal temperature *** results may provide information for forest protection and ecological construction in this region,and a scientific reference for future climate reconstruction.

关键词： Tree ring width Lubanling Pinus armandii Franch Slope direction Climate response

同方期刊数据库

L vector spaces and L fields

在线全文

学校读者我要写书评

暂无评论

Science China Mathematics 2024年第10期67卷 2195-2216页

作者： Yinhe peng Liuzhen Wu Academy of Mathematics and Systems Science Chinese Academy of SciencesBeijing 100190China

We construct in ZFC(the Zermelo-Fraenkel system with choice) an L topological vector space—a topological vector space that is an L space—and an L field—a topological field that is an L space. This generalizes earli... 详细信息

关键词： minimal walk topological vector space topological eld L space L vector space L eld

Semi-derived Ringel-Hall algebras and Hall algebras of odd-periodic relative derived categories

在线全文

学校读者我要写书评

暂无评论

Science China Mathematics 2024年第8期67卷 1735-1760页

作者： Ji Lin Liangang peng Department of Mathematics and Statistics Fuyang Normal UniversityFuyang 236037China School of Mathematics Sichuan UniversityChengdu 610064China

Let t be a positive integer and A be a hereditary abelian category satisfying some finiteness *** define the semi-derived Ringel-Hall algebra of A from the category C_(Z/t)(A)of Z/t-graded complexes and obtain a natural basis of the semi-derived Ringel-Hall ***,we describe the semiderived Ringel-Hall algebra by the generators and defining *** particular,if t is an odd integer,we show an embedding of the derived Hall algebra of the odd-periodic relative derived category in the extended semi-derived Ringel-Hall algebra.

关键词： hereditary abelian categories the categories of Z/t-graded complexes semi-derived Ringel-Hall algebras derived Hall algebras