中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 视频内容安全团队
Tangent Fisher Vector on Matrix Manifolds for Action Recognition

文献类型：期刊论文

作者Luo, Guan1; Wei, Jiutong1; Hu, Weiming1; Maybank, Stephen J.2
刊名IEEE TRANSACTIONS ON IMAGE PROCESSING
出版日期2020
卷号29页码:3052-3064
关键词Manifolds Video sequences Observability Videos Covariance matrices Kernel Computational modeling Action recognition Fisher vector Grassmann manifold Hankel matrix matrix manifold
ISSN号1057-7149
DOI10.1109/TIP.2019.2955561
通讯作者Hu, Weiming(wmhu@nlpr.ia.ac.cn)
英文摘要In this paper, we address the problem of representing and recognizing human actions from videos on matrix manifolds. For this purpose, we propose a new vector representation method, named tangent Fisher vector, to describe video sequences in the Fisher kernel framework. We first extract dense curved spatio-temporal cuboids from each video sequence. Compared with the traditional 'straight cuboids', the dense curved spatio-temporal cuboids contain much more local motion information. Each cuboid is then described using a linear dynamical system (LDS) to simultaneously capture the local appearance and dynamics. Furthermore, a simple yet efficient algorithm is proposed to learn the LDS parameters and approximate the observability matrix at the same time. Each video sequence is thus represented by a set of LDSs. Considering that each LDS can be viewed as a point in a Grassmann manifold, we propose to learn an intrinsic GMM on the manifold to cluster the LDS points. Finally a tangent Fisher vector is computed by first accumulating all the tangent vectors in each Gaussian component, and then concatenating the normalized results across all the Gaussian components. A kernel is defined to measure the similarity between tangent Fisher vectors for classification and recognition of a video sequence. This approach is evaluated on the state-of-the-art human action benchmark datasets. The recognition performance is competitive when compared with current state-of-the-art results.
WOS关键词BINET-CAUCHY KERNELS ; DYNAMICAL-SYSTEMS ; VIEW ; MODELS ; VIDEO ; IDENTIFICATION ; CLASSIFICATION ; DESCRIPTORS
资助项目NSFC-General Technology Collaborative Fund for Basic Research[U1636218] ; Natural Science Foundation of China[61751212] ; Natural Science Foundation of China[61721004] ; Beijing Natural Science Foundation[L172051] ; Key Research Program of Frontier Sciences, CAS[QYZDJSSW-JSC040] ; CAS External Cooperation Key Project ; National Natural Science Foundation of Guangdong[2018B030311046]
WOS研究方向Computer Science ; Engineering
语种英语
WOS记录号WOS:000510750900013
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
资助机构NSFC-General Technology Collaborative Fund for Basic Research ; Natural Science Foundation of China ; Beijing Natural Science Foundation ; Key Research Program of Frontier Sciences, CAS ; CAS External Cooperation Key Project ; National Natural Science Foundation of Guangdong
源URL[http://ir.ia.ac.cn/handle/173211/38369]  
专题自动化研究所_模式识别国家重点实验室_视频内容安全团队
通讯作者Hu, Weiming
作者单位1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
2.Birkbeck Coll, Dept Comp Sci & Informat Syst, London WC1E 7HX, England
推荐引用方式
GB/T 7714		 Luo, Guan,Wei, Jiutong,Hu, Weiming,et al. Tangent Fisher Vector on Matrix Manifolds for Action Recognition[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2020,29:3052-3064.
APA Luo, Guan,Wei, Jiutong,Hu, Weiming,&Maybank, Stephen J..(2020).Tangent Fisher Vector on Matrix Manifolds for Action Recognition.IEEE TRANSACTIONS ON IMAGE PROCESSING,29,3052-3064.
MLA Luo, Guan,et al."Tangent Fisher Vector on Matrix Manifolds for Action Recognition".IEEE TRANSACTIONS ON IMAGE PROCESSING 29(2020):3052-3064.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。