CASIA OpenIR  > 视频内容安全团队
Asymmetric 3D Convolutional Neural Networks for action recognition
Yang, Hao1,3; Yuan, Chunfeng1; Li, Bing1; Du, Yang1,3; Xing, Junliang1; Hu, Weiming1,2,3; Maybank, Stephen J.4
Source PublicationPATTERN RECOGNITION
ISSN0031-3203
2019
Volume85Issue:1Pages:1-12
Abstract

Convolutional Neural Network based action recognition methods have achieved significant improvements in recent years. The 3D convolution extends the 2D convolution to the spatial-temporal domain for better analysis of human activities in videos. The 3D convolution, however, involves many more parameters than the 2D convolution. Thus, it is much more expensive on computation, costly on storage, and difficult to learn. This work proposes efficient asymmetric one-directional 3D convolutions to approximate the traditional 3D convolution. To improve the feature learning capacity of asymmetric 3D convolutions, a set of local 3D convolutional networks, called MicroNets, are proposed by incorporating multi-scale 3D convolution branches. Then, an asymmetric 3D-CNN deep model is constructed by MicroNets for the action recognition task. Moreover, to avoid training two networks on the RGB and Flow frames separately as most works do, a simple but effective multi-source enhanced input is proposed, which fuses useful information of the RGB and Flow frame at the pre-processing stage. The asymmetric 3D-CNN model is evaluated on two of the most challenging action recognition benchmarks, UCF-101 and HMDB-51. The asymmetric 3D-CNN model outperforms all the traditional 3D-CNN models in both effectiveness and efficiency, and its performance is comparable with that of recent state-of-the-art action recognition methods on both benchmarks. (C) 2018 Elsevier Ltd. All rights reserved.

KeywordAsymmetric 3D convolution MicroNets 3D-CNN Action recognition
DOI10.1016/j.patcog.2018.07.028
WOS KeywordSHORT-TERM-MEMORY ; FEATURES ; VIDEOS ; TIME ; FLOW
Indexed BySCI
Language英语
WOS Research AreaComputer Science ; Engineering
WOS SubjectComputer Science, Artificial Intelligence ; Engineering, Electrical & Electronic
WOS IDWOS:000447819300001
PublisherELSEVIER SCI LTD
Citation statistics
Document Type期刊论文
Identifierhttp://ir.ia.ac.cn/handle/173211/22816
Collection视频内容安全团队
Corresponding AuthorYuan, Chunfeng
Affiliation1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
2.Chinese Acad Sci, CAS Ctr Excellence Brain Sci & Intelligence Techn, Beijing 100190, Peoples R China
3.Univ Chinese Acad Sci, Beijing 100190, Peoples R China
4.Birkbeck Coll, Dept Comp Sci & Informat Syst, London WC1E 7HX, England
First Author AffilicationChinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
Corresponding Author AffilicationChinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
Recommended Citation
GB/T 7714
Yang, Hao,Yuan, Chunfeng,Li, Bing,et al. Asymmetric 3D Convolutional Neural Networks for action recognition[J]. PATTERN RECOGNITION,2019,85(1):1-12.
APA Yang, Hao.,Yuan, Chunfeng.,Li, Bing.,Du, Yang.,Xing, Junliang.,...&Maybank, Stephen J..(2019).Asymmetric 3D Convolutional Neural Networks for action recognition.PATTERN RECOGNITION,85(1),1-12.
MLA Yang, Hao,et al."Asymmetric 3D Convolutional Neural Networks for action recognition".PATTERN RECOGNITION 85.1(2019):1-12.
Files in This Item: Download All
File Name/Size DocType Version Access License
Asymmetric 3D CNN 发表(2689KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Download
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Yang, Hao]'s Articles
[Yuan, Chunfeng]'s Articles
[Li, Bing]'s Articles
Baidu academic
Similar articles in Baidu academic
[Yang, Hao]'s Articles
[Yuan, Chunfeng]'s Articles
[Li, Bing]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Yang, Hao]'s Articles
[Yuan, Chunfeng]'s Articles
[Li, Bing]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: Asymmetric 3D CNN 发表版.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.