检索结果-南通市图书馆

Multi‐granularity re‐ranking for visible‐infrared person 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 2023年第3期8卷 770-779页

作者： Yadi Wang Hongyun Zhang duoqian miao Witold Pedrycz Department of Computer Science and Technology Tongji UniversityShanghaiChina Key Laboratory of Embedded System and Service Computing Ministry of EducationShanghaiChina Department of Electrical and Computer Engineering University of AlbertaEdmontonAlbertaCanada System Research Institute Polish Academy of SciencesWarsawPoland

Visible‐infrared person re‐identification(VI‐ReID)is a supplementary task of single‐modality re‐identification,which makes up for the defect of conventional re‐identification under insufficient illumination.It is more challenging than single‐modality ReID because,in addition to difficulties in pedestrian posture,camera shoot-ing angle and background change,there are also difficulties in the cross‐modality gap.Existing works only involve coarse‐grained global features in the re‐ranking calculation,which cannot effectively use fine‐grained features.However,fine‐grained features are particularly important due to the lack of information in cross‐modality re‐ID.To this end,the Q‐center Multi‐granularity K‐reciprocal Re‐ranking Algorithm(termed QCMR)is proposed,including a Q‐nearest neighbour centre encoder(termed QNC)and a Multi‐granularity K‐reciprocal Encoder(termed MGK)for a more comprehensive feature representation.QNC converts the probe‐corresponding modality features into gallery corresponding modality features through modality transfer to narrow the modality gap.MGK takes a coarse‐grained mutual nearest neighbour as the dominant and combines a fine‐grained nearest neighbour as a supplement for similarity measurement.Extensive experiments on two widely used VI‐ReID benchmarks,SYSU‐MM01 and RegDB have shown that our method achieves state‐of‐the‐art results.Especially,the mAP of SYSU‐MM01 is increased by 5.9%in all‐search mode.

关键词： computer vision recognition

Influence of kernel clustering on an RBFN

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 2019年第4期4卷 255-260页

作者： Changming Zhu duoqian miao College of Information Engineering Shanghai Maritime UniversityNo.1550Haigang AvenueShanghaiPeople’s Republic of China Department of Computer Science and Technology Tongji UniversityNo.4800Cao’an HighwayShanghaiPeople’s Republic of China

Classical radial basis function network(RBFN)is widely used to process the non-linear separable data sets with the introduction of activation functions.However,the setting of parameters for activation functions is random and the distribution of patterns is not taken into account.To process this issue,some scholars introduce the kernel clustering into the RBFN so that the clustering results are related to the parameters about activation functions.On the base of the original kernel clustering,this study further discusses the influence of kernel clustering on an RBFN when the setting of kernel clustering is changing.The changing involves different kernel-clustering ways[bubble sort(BS)and escape nearest outlier(ENO)],multiple kernel-clustering criteria(static and dynamic)etc.Experimental results validate that with the consideration of distribution of patterns and the changes of setting of kernel clustering,the performance of an RBFN is improved and is more feasible for corresponding data sets.Moreover,though BS always costs more time than ENO,it still brings more feasible clustering results.Furthermore,dynamic criterion always cost much more time than static one,but kernel number derived from dynamic criterion is fewer than the one from static.

关键词： kernel RBFN activation

Constrained tolerance rough set in incomplete information systems

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4期6卷 440-449页

作者： Renxia Wan duoqian miao Witold Pedrycz School of Mathematics and Inforrmation Scicncc North Minzu UniversityYinchuanNingxiaChina Hongyang Institute for Big Data in Health FuzhouFujianChina Department of Computer Science and Technology Tongi UniversityShanghaiChina Department of Electrical and Computer Engineering University of AlbertaEdmontonCanada

The tolerance rough set is developed as one of the outstanding extensions of the Pawlak's rough set model under incomplete information,and the limited tolerance relation is developed to overcome the problem that objects leniently satisfy the tolerance relation.However,the classification based on the limited tolerance relationship cannot reflect the matching degree of uncertain information of objects.In this article,we explore the influence of null values in an incomplete system,and propose the constrained tolerance relation based on the matching degree of uncertain information of objects.The proposed rough set based on the constrained tolerance relation can provide a more detailed structure of an object class through threshold.Proofs and example analyses further show the rationality and superiority of the proposed model.

关键词： incomplete matching constrained

Uncertainty-optimized deep learning model for small-scale person 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

Science China(Information Sciences) 2019年第12期62卷 20-32页

作者： Cairong ZHAO Kang CHEN Di ZANG Zhaoxiang ZHANG Wangmeng ZUO duoqian miao Department of Computer Science and Technology Tongji University Institute of Automation Chinese Academy of Sciences School of Computer Science and Technology Harbin Institute of Technology

In recent years, deep learning has developed rapidly and is widely used in various fields, such as computer vision, speech recognition, and natural language processing. For end-to-end person re-identification,most deep learning methods rely on large-scale datasets. Relatively few methods work with small-scale datasets. Insufficient training samples will affect neural network accuracy significantly. This problem limits the practical application of person re-identification. For small-scale person re-identification, the uncertainty of person representation and the overfitting problem associated with deep learning remain to be solved.Quantifying the uncertainty is difficult owing to complex network structures and the large number of hyperparameters. In this study, we consider the uncertainty of pedestrian representation for small-scale person re-identification. To reduce the impact of uncertain person representations, we transform parameters into distributions and conduct multiple sampling by using multilevel dropout in a testing process. We design an improved Monte Carlo strategy that considers both the average distance and shortest distance for matching and ranking. When compared with state-of-the-art methods, the proposed method significantly improve accuracy on two small-scale person re-identification datasets and is robust on four large-scale datasets.

关键词： person re-identification uncertainty analysis deep learning

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

一种多任务联合训练的机器阅读理解模型

小型微型计算机系统 2024年第6期45卷 1398-1404页

作者：王勇陈秋怡苗夺谦杨宁创重庆理工大学两江人工智能学院重庆401135 同济大学电子与信息工程学院上海201804

在机器阅读理解任务中,如何在包含不可回答问题的情况下提高答案的准确性是自然语言处理领域的一项重要挑战.虽然基于深度学习的机器阅读理解模型展现出很好的性能,但是这些模型仍然存在抽取特征冗余、语义信息不全面、问题分类任务和... 详细信息

在机器阅读理解任务中,如何在包含不可回答问题的情况下提高答案的准确性是自然语言处理领域的一项重要挑战.虽然基于深度学习的机器阅读理解模型展现出很好的性能,但是这些模型仍然存在抽取特征冗余、语义信息不全面、问题分类任务和答案抽取任务耦合性不强的问题.为了解决以上问题,本文提出一种结合门控机制和多级残差结构的多任务联合训练模型GMRT(Gated Mechanism and Multi-level Residual Structure for Multi-task Joint Training),以提升机器阅读理解任务中答案预测的准确性.GMRT构建门控机制来筛选交互后的关联特征,从而控制信息的流动.采用多级残差结构分别连接注意力机制和门控机制,保证每个阶段都保留原始语义信息.同时,通过边缘损失函数对问题分类任务和答案抽取任务联合训练,确保预测答案过程中任务之间的强耦合性.在SQuAD2.0数据集上的实验结果表明,GMRT模型的EM值和F1值均优于对比模型.

关键词：机器阅读理解多任务联合训练门控机制多级残差结构

基于簇中心预选策略的三支决策密度峰值聚类算法

维普期刊数据库

同方期刊数据库博看期刊评论

在线全文

学校读者我要写书评

暂无评论

山西大学学报(自然科学版) 2024年第1期47卷 30-39页

作者：罗舒文万仁霞苗夺谦泉州信息工程学院通识教育中心北方民族大学数学与信息科学学院同济大学电子与信息工程学院

本文针对密度峰值聚类算法（CFSFDP）无法自动选取簇中心的不确定性问题，通过引入三支决策理论对其进行优化，提出了一种基于簇中心预选策略的三支决策密度峰值聚类算法（TDPC）。首先利用密度和距离两参数的统计特性将数据对象划分核... 详细信息

本文针对密度峰值聚类算法（CFSFDP）无法自动选取簇中心的不确定性问题，通过引入三支决策理论对其进行优化，提出了一种基于簇中心预选策略的三支决策密度峰值聚类算法（TDPC）。首先利用密度和距离两参数的统计特性将数据对象划分核心域、边界域与琐碎域，符合条件的聚类中心被置于核心域，难以判定的疑似聚类中心点则被置于边界域，然后通过定义的k-可达域和判别准则对疑似聚类中心进行分析，选取出实际聚类中心。所提出算法有效解决了密度峰值聚类算法聚类中心自动确定问题。在2个人工数据集和4个UCI(University of California, lrvine)公共数据集上对TDPC进行测试。与CFSFDP算法和DBSCAN(Density-Based Spatial Clustering of Applications with Noise)算法进行聚类性能比较，所提出算法TDPC在轮廓系数、DB(Davies-Bouldin)指数、调整互信息、调整兰德系数、FM(Fowlkes-Mallows)指数、同质性、完整性等聚类评价指标方面均达到最优或与最优算法结果相近，表明TDPC综合聚类性能优于比较算法，具有良好的聚类可行性与有效性。

关键词：聚类算法聚类中心边界域三支聚类密度聚类 k-可达域

Guest Editorial: Rough Sets and Data Mining

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

CAAI Transactions on Intelligence Technology 2019年第4期4卷 201-202页

作者： HIROSHI SAKAI MICHINORI NAKATA WEI-ZHI WU duoqian miao GUOYIN WANG Faculty of Engineering Kyushu Institute of TechnologySensuiTobataKitakyushu 804-8550Japan Faculty of Management and Information Science Josai International UniversityToganeChiba 283-0002Japan School of Mathematics Physics and Information ScienceZhejiang Ocean UniversityZhoushanZhejiang316022China Department of Computer Science and Technology Tongji UniversityShanghai201804China Chongqing Key Laboratory of Computational Intelligence Chongqing University of Posts and TelecommunicationsChongqingChina

A rough set,first described by Polish computer scientist Zdzis?aw Pawlak,is a formal approximation of a crisp set,and it is now known as a new mathematical tool to process vague concepts.They are used for machine learning,knowledge discovery,feature selection,etc.,and are applied to artificial intelligence,medical informatics,civil engineering,Kansei engineering,decision science,business administration,and so on.Especially,research on data mining using rough sets is widely spreading,and the obtained association rules are applied to the characterisation of data and decision support.

关键词： Rough artificial computer

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

基于三支决策的蚁群聚类算法

昆明理工大学学报（自然科学版） 2024年第1期49卷 83-97页

作者：王梦绚万仁霞苗夺谦赵杰北方民族大学计算机科学与工程学院宁夏银川750021 同济大学电子与信息工程学院上海201804

针对蚁群聚类算法在蚂蚁之间缺少信息交互导致误识别率高和蚂蚁单独移动带来的资源浪费的问题,本文将三支决策思想融入蚁群聚类算法来改进其性能.通过蚂蚁激活机制、微簇生成机制将待聚类的蚂蚁“一分为三”,区分为已激活的单个蚂蚁、... 详细信息

针对蚁群聚类算法在蚂蚁之间缺少信息交互导致误识别率高和蚂蚁单独移动带来的资源浪费的问题,本文将三支决策思想融入蚁群聚类算法来改进其性能.通过蚂蚁激活机制、微簇生成机制将待聚类的蚂蚁“一分为三”,区分为已激活的单个蚂蚁、未激活的单个蚂蚁和蚂蚁微簇.再使用精英蚂蚁机制、二次验证机制,遴选相似度达到一定阈值的微簇,形成聚类正域,并赋予精英蚂蚁更高的优先级和固定的平面位置,最后利用不同正域间的信息熵值为属性加权,引导边界域中蚂蚁向着更相似且优先级更高的蚂蚁方向移动.实验结果表明,本文所提出的算法不仅提升了蚁群聚类的质量,还具有良好的时间效率.

关键词：蚁群聚类算法三支决策精英蚂蚁微簇属性加权

维普期刊数据库

在线全文

学校读者我要写书评

暂无评论

基于多粒度空间混乱的细粒度图像分类算法

智能系统学报 2022年第1期17卷 144-150页

作者：宋思雨苗夺谦同济大学电子与信息工程学院上海201804 同济大学嵌入式系统与服务计算教育部重点实验室上海201804

细粒度图像分类是计算机视觉领域一个具有挑战性的任务,在实际场景中具有很高的应用价值。其中不同子类别的物体整体轮廓差异较小,微小的判别性局部区域是分类的关键。然而,这些重要的局部区域的尺度可能不同,不能用单一的标准去衡量它... 详细信息

细粒度图像分类是计算机视觉领域一个具有挑战性的任务,在实际场景中具有很高的应用价值。其中不同子类别的物体整体轮廓差异较小,微小的判别性局部区域是分类的关键。然而,这些重要的局部区域的尺度可能不同,不能用单一的标准去衡量它们。为了解决这个问题,本文提出了多粒度空间混乱模块来帮助神经网络学习如何寻找到不同尺度的判别性细节。该模块首先将图片划分为不同粒度的局部区域,然后随机打乱并重组构成新的输入图片。经过处理的图片具有区域无关性,从而迫使网络更好地在不同粒度层次下寻找有判别力的局部区域并从中学习特征。在3个广泛使用的细粒度图像分类数据集上的实验证明本文提出的模块可以有效地帮助网络寻找判别性局部区域从而提升了准确率并且网络不需要图片的任何部位标注信息。

关键词：细粒度图像分类空间混乱多粒度深度学习数据增强卷积神经网络弱监督学习局部区域

维普期刊数据库