检索结果-南通市图书馆

Asymmetric Deep Hashing for Person Re-Identifications

在线全文

学校读者我要写书评

暂无评论

Tsinghua Science and Technology 2022年第2期27卷 396-411页

作者： yali Zhao yali Li Shengjin wang Department of Electronic Engineering Tsinghua UniversityBeijing 100084China

The person re-identification(re-ID)community has witnessed an explosion in the scale of data that it has to handle.On one hand,it is important for large-scale re-ID to provide constant or sublinear search time and dramatically reduce the storage cost for data points from the viewpoint of efficiency.On the other hand,the semantic affinity existing in the original space should be preserved because it greatly boosts the accuracy of re-ID.To this end,we use the deep hashing method,which utilizes the pairwise similarity and classification label to learn deep hash mapping functions,in order to provide discriminative representations.More importantly,considering the great advantage of asymmetric hashing over the existing symmetric one,we finally propose an asymmetric deep hashing(ADH)method for large-scale re-ID.Specifically,a two-stream asymmetric convolutional neural network is constructed to learn the similarity between image pairs.Another asymmetric pairwise loss is formulated to capture the similarity between the binary hashing codes and real-value representations derived from the deep hash mapping functions,so as to constrain the binary hash codes in the Hamming space to preserve the semantic structure existing in the original space.Then,the image labels are further explored to have a direct impact on the hash function learning through a classification loss.Furthermore,an efficient alternating algorithm is elaborately designed to jointly optimize the asymmetric deep hash functions and high-quality binary codes,by optimizing one parameter with the other parameters fixed.Experiments on the four benchmarks,i.e.,DukeMTMC-reID,Market-1501,Market-1501+500 k,and CUHK03 substantiate the competitive accuracy and superior efficiency of the proposed ADH over the compared state-of-the-art methods for large-scale re-ID.

关键词： person re-identification deep hashing asymmetric hashing large-scale

Towards robustness and generalization of point cloud representation:A geometry coding method and a large-scale object-level dataset

在线全文

学校读者我要写书评

暂无评论

Computational Visual Media 2024年第1期10卷 27-43页

作者： Mingye Xu Zhipeng Zhou Yali Wang Yu Qiao Guangdong-Hong Kong-Macao Joint Laboratory of Human-Machine Intelligence-Synergy Systems Shenzhen Institute of Advanced TechnologyChinese Academy of SciencesShenzhen518000China University of Chinese Academy of Sciences Beijing065001China Shanghai AI Laboratory Shanghai200001China SIAT Branch Shenzhen Institute of Artificial Intelligence and Robotics for SocietyShenzhen518000China Alibaba DAMO Academy Hangzhou242332China

Robustness and generalization are two challenging problems for learning point cloud representation.To tackle these problems,we first design a novel geometry coding model,which can effectively use an invariant eigengraph to group points with similar geometric information,even when such points are far from each other.We also introduce a large-scale point cloud dataset,PCNet184.It consists of 184 categories and 51,915 synthetic objects,which brings new challenges for point cloud classification,and provides a new benchmark to assess point cloud cross-domain generalization.Finally,we perform extensive experiments on point cloud classification,using ModelNet40,ScanObjectNN,and our PCNet184,and segmentation,using ShapeNetPart and S3DIS.Our method achieves comparable performance to state-of-the-art methods on these datasets,for both supervised and unsupervised learning.Code and our dataset are available at https://***/MingyeXu/PCNet184.

关键词： geometry coding self-supervised learning point cloud classification segmentation 3D analysis

Spatial and temporal dynamics of microbial community composition and factors influencing the surface water and sediments of urban rivers

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

Journal of Environmental Sciences 2023年第2期124卷 187-197页

作者： Hongjie wang Xingchun Liu Yali Wang Shengqi Zhang Guangming Zhang Yangyang Han Mengxiang Li Ling Liu College of Ecology and Environment Hebei UniversityBaoding 071002China Institute of Life Science and Green Development Hebei UniversityBaoding 071002China Hebei Key Laboratory of Close-to-Nature Restoration Technology of Wetlands Baoding 071002China School of Energy and Environmental Engineering Hebei University of TechnologyTianjin 300130China

The temporal and spatial characteristics of urban river bacterial communities help us understand the feedback mechanism of bacteria to changes in the aquatic environment.The Fuhe River plays an important role in determining the water ecological environment of Baiyangdian Lake.16S rRNA gene sequencing was used to study the microbial distribution characteristics in the Fuhe River in different seasons.The results showed that some environmental factors of the surface water(ammonia nitrogen(NH_(3)^(-)N),total nitrogen(TN),and total phosphorus(TP))were different on the spatial and temporal scales.Moreover,there were no seasonal differences in the contents of TN,TP,total organic carbon(TOC),or heavy metals in the sediments.The distributions of Cyanobacteria,Actinomycetes and Firmicutes in the water and Actinomycetes and Planctomycetes in the sediments differed significantly among seasons(P<0.05).There were significant spatial differences in bacteria in the surface water,with the highest abundance of Proteobacteria recorded in the river along with the highest nutrient concentration,while the abundance of Bacteroidetes was higher in the upstream than the downstream.Microbial communities in the water weremost sensitive to temperature(T)and the TP concentration(P<0.01).Moreover,differences in the bacterial community were better explained by the content of heavy metals in the sediments than by the chemical characteristics.A PICRUStmetabolic inference analysis showed that the effect of high summer temperatures on the enzyme action led to an increase in the abundances of the metabolic-related genes of the river microorganisms.

关键词： Fuhe river Surface water Sediment Bacterial community Spatial and temporal distribution Functional prediction

Person Re-Identification with Effectively Designed Parts

在线全文

学校读者我要写书评

暂无评论

Tsinghua Science and Technology 2020年第3期25卷 415-424页

作者： yali Zhao yali Li Shengjin wang Tsinghua University

Person re-IDentification(re-ID) is an important research topic in the computer vision community, with significance for a range of applications. Pedestrians are well-structured objects that can be partitioned, although detection errors cause slightly misaligned bounding boxes, which lead to mismatches. In this paper, we study the person re-identification performance of using variously designed pedestrian parts instead of the horizontal partitioning routine typically applied in previous hand-crafted part works, and thereby obtain more effective feature descriptors. Specifically, we benchmark the accuracy of individual part matching with discriminatively trained Convolutional Neural Network(CNN) descriptors on the Market-1501 dataset. We also investigate the complementarity among different parts using combination and ablation studies, and provide novel insights into this issue. Compared with the state-of-the-art, our method yields a competitive accuracy rate when the best part combination is used on two large-scale datasets(Market-1501 and CUHK03) and one small-scale dataset(VIPeR).

关键词： person re-IDentification(re-ID) Convolutional Neural Network(CNN) part model

Molecular phylogeny and historical biogeography of the cave fish genus Sinocyclocheilus(Cypriniformes:Cyprinidae)in southwest China

在线全文

学校读者我要写书评

暂无评论

Integrative Zoology 2022年第2期17卷 311-325页

作者： Huamei WEN Tao LUO Yali Wang Siwei wang Tao LIU Ning XIAO Jiang ZHOU School of Life Sciences Central China Normal UniversityWuhanChina School of Karst Sciences Guizhou Normal UniversityGuiyangChina School of Life Sciences Guizhou Normal UniversityGuiyangChina Department of Senior High School Liupanshui No.4 Higth SchoolLiupanshuiChina Department of Basic Medicine Guiyang Nursing Vocational CollegeGuiyangChina

Modern accumulations of genetic data offer unprecedented opportunities for understanding the systematic classification and origins of specific groups of organisms.The genus Sinocyclocheilus is among the most cave abundant genera in Cyprinidae,with 76 recognized species,belonging to 4 species groups.Recent phylogenetic studies have shown that the classification of species groups within the genus Sinocyclocheilus remains controversial.In this study,we constructed a sequence supermatrix of 26 species from 4 species groups of the genus Sinocyclocheilus using the mitochondrial genome to reveal phylogenetic relationships,historical biogeography and patterns of species diversification in the genus Sinocyclocheilus.Phylogenetic analysis strongly supports the monophyletic groups of the 3 species groups(S.jii,S.cyphotergous,and S.tingi groups)except the S.angularis group.Phylogenetic analysis showed that S.anshuiensis and S.microphthalmus,which were recognized as numbers of S.angularis group,formed a strongly supported independent clade.Therefore,we propose a new species group,the S.microphthalmus group,which contains S.anshuiensis and S.microphthalmus.Biogeographic reconstruction suggests that the living Sinocyclocheilus may have originated in north-central Guangxi at the late Eocene and dispersed outward after a vicariance at 32.31 Million years ago(Ma).Early diversification is focused on the late Oligocene(ca.25 Ma),which is related to the second uplift of the Qinghai−Tibetan Plateau and the lateral extrusion of the Indochina at the Oligocene/Miocene boundary.Our results suggest that 2 uplifts of the Qinghai−Tibetan Plateau and climate change in the Miocene may have influenced the diversification of the Sinocyclocheilus lineage.

关键词： biogeography cave fish karst phylogeny Sinocyclocheilus

Failure prognostic of systems with hidden degradation process

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

Journal of Systems Engineering and Electronics 2012年第2期23卷 314-324页

作者： Yali Wang Wenhai wang Department of Control Science and Engineering Zhejiang UniversityHangzhou 310027R R.China

Systems with a hidden degradation process are perva- sive in the real world. Degrading critical components will under- mine system performance and pose potential failures in the future. Prognostic aims at predicting potential failures before it evolves into faults. A prognostic procedure based on expectation maxi- mization and unscented Kalman filter is proposed. System state, sensor measurement and hidden degradation process are viewed as data （incomplete or missing） in the expectation maximization method. System state and hidden degradation process are esti- mated by a unscented Kalman filter upon sensor measurements. Component-specific parameters in a degradation process are iden- tified on the estimation of the degradation process. Residual life is measured by the median of estimated residual life distribution. The proposed procedure is verified by simulations on a first-order capacitor-resistance circuit with degrading resistance. Residual life estimation consists conservatively with the trend and is evalu- ated in terms of relative errors. Simulation results are reasonable. The proposed prognostic method expects applications in practice.

关键词： prognostic degradation expectation maximization unscented Kalman filter （UKF） residual life.

Edge atomic Fe sites decorated porous graphitic carbon as an efficient bifunctional oxygen catalyst for Zinc-air batteries

在线全文

学校读者我要写书评

暂无评论

能源化学:英文版 2023年第8期83卷 602-611,I0014页

作者： Ruihui Gan Yali Wang Xiangwu Zhang Yan Song Jingli Shi Chang Ma Tianjin Municipal Key Lab of Advanced Fiber and Energy Storage Technology Tiangong UniversityTianjin 300387China Fiber and Polymer Science Program Department of Textile EngineeringChemistry and ScienceWilson College of TextilesNorth Carolina State UniversityRaleighNC 27695-8301uNITED sTATES CAS Key Laboratory of Carbon Materials Institute of Coal ChemistryChinese Academy of SciencesTaiyuan 030001ShanxiChina

The development of advanced bifunctional oxygen electrocatalysts for oxygen reduction and evolution reactions(ORR and OER) is critical to the practical application of zinc-air batteries(ZABs). Herein, a silica-assisted method is reported to integrate numerous accessible edge Fe-Nx sites into porous graphitic carbon(named Fe-N-G) for achieving highly active and robust oxygen electrocatalysis. Silica facilitates the formation of edge Fe-Nx sites and dense graphitic domains in carbon by inhibiting iron aggregation.The purification process creates a well-developed mass transfer channel for Fe-N-G. Consequently,Fe-N-G delivers a half-wave potential of 0.859 V in ORR and an overpotential of 344 m V at10 m A cm^(-2)in OER. During long-term operation, the graphitic layers protect edge Fe-Nx sites from demetallation in ORR and synergize with Fe OOH species endowing Fe-N-G with enhanced OER activity.Density functional theory calculations reveal that the edge Fe-Nx site is superior to the in-plane Fe-Nx site in terms of OH* dissociation in ORR and OOH* formation in OER. The constructed ZAB based on Fe-N-G cathode shows a higher peak power density of 133 m W cm^(-2)and more stable cycling performance than Pt/C + RuO2counterparts. This work provides a novel strategy to obtain high-efficiency bifunctional oxygen electrocatalysts through space mediation.

关键词： Bifunctionality EdgeFe-Nxsites Oxygen catalysis Graphitic domains Zinc-air batteries

Providing Delay-Guaranteed Multicast Transmission Services for an SDN-Based Publish/Subscribe System

在线全文

学校读者我要写书评

暂无评论

China Communications 2019年第12期16卷 176-187页

作者： Yali Wang Shuangxi Hong Junna Zhang College of Computer and Information Engineering Henan Normal UniversityXinxiang 450002China Engineering Laboratory of Intellectual Business and Internet of Things Technologies Henan ProvinceChina

Providing end-to-end delay guarantees in traditional Internet is a complex task due to the distributed nature of TCP/IP protocols. Software Defined Networking(SDN) gives a new dimension to improve QoS(Quality of Service) as it can benefit from its flexibility, programmability and centralized view. In this paper, we provide delay-guaranteed data transmission service instead of "best efforts" service for a topic-based publish/subscribe system by means of exploring these specific features of SDN. We attribute this routing problem in such conditions to Delay-Constraint Lowest Cost Steiner Tree(DCLCST) problem. To solve it, we compute the shortest delay paths from source node to every subscribe node and the shortest cost paths from every subscribe node to any other node using dijkstra algorithm. Then we construct a delay-constraint least cost steiner tree for per-topic based on these paths as multicast tree. We also present experimental results to demonstrate the effectiveness of the algorithms and methods we proposed.

关键词： minimum steiner tree multicast routing publish/subscribe system overlay network

Vertical distribution characteristics of soil mercury and its formation mechanism in permafrost regions:A case study of the Qinghai-Tibetan Plateau

在线全文

学校读者我要写书评

暂无评论

Journal of Environmental Sciences 2022年第3期34卷 311-321页

作者： yali Liu Junfeng wang Junming Guo Luyang wang Qingbai Wu State Key Laboratory of Frozen Soil Engineering Northwest Institute of Eco-Environment and ResourcesChinese Academy of SciencesLanzhou 730000China University of Chinese Academy of Sciences Bering 200049China State Key Laboratory of Cryospheric Science Northwest Institute of Eco-environment and ResourcesChinese Academy of Sciences(CAS)Lanzhou 730000China

Continuing permafrost degradation is increasing the risk of mercury(Hg) exposure in the permafrost regions on the Qinghai-Tibetan Plateau(QTP),but related studies are still limited,especially the ones on the detailed Hg migration processes in permafrost.The vertical distribution characteristics of soil Hg were investigated in three ecosystems in the Beiluhe area on the QTP,and its influencing factors and formation mechanism were investigated.The results indicate that the total soil mercury(THg)concentration in the Beiluhe area remains at an extremely low level(6.33±2.45 ng/g).In the vertical profile,the THg concentration of the shallow soil layer(0-50 cm)(5.96±2.22 ng/g)is significantly lower than that of the deep layer(50-400 cm)(7.44±2.71 ng/g)(p<0.05).Within the upper 50 cm,the THg concentration decreases with soil depth,and the peak THg concentration occurs at 100-300 cm on the entire profile.Although the THg concentration is slightly affected by the organic matter in the shallow soil layer,in general,the soil parent material is the dominant factor affecting the THg concentration.Intense weathering results in a low THg concentration in the shallow soil layer because the soil Hg is carried downward with the soil moisture.To a certain depth,the impermeable frozen soil layer intercepts the flow of the soil Hg,and it forms a Hg enrichment layer.This paper presents the distinctive pattern of the soil Hg distribution in the permafrost regions of the QTP.

关键词： Total soil mercury Qinghai-Tibetan Plateau Alpine ecosystems Permafrost Mercury migration