Video action recognition method based on attention residual network and LSTM

Yu Zhang; Pengyue Dong

doi:10.1109/CCDC52312.2021.9601577

Video action recognition method based on attention residual network and LSTM

Yu Zhang^*, Pengyue Dong

^*Corresponding author for this work

Beijing University of Civil Engineering and Architecture

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Citation (Scopus)

Abstract

A video action recognition method based on attention residual network and long-term memory network(LSTM) is proposed, which is to solve the problems that the existing human action recognition methods are prone to overfitting, susceptible to interference information, and lack of feature expression ability. In the beginning, the traditional data preprocessing method and sampling method are improved to enhance the generalization ability of the model. Then, a residual network with attention is proposed to improve the feature extraction ability of the network. At length, LSTM is used to recognize video actions. Experimental results on UCF YouTube dataset show that the proposed method can recognize the actions in video more effectively than other similar methods in this field, and the recognition rate reaches 95.45%.

Original language	English
Title of host publication	Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	3611-3616
Number of pages	6
ISBN (Electronic)	9781665440899
DOIs	https://doi.org/10.1109/CCDC52312.2021.9601577
Publication status	Published - 2021
Externally published	Yes
Event	33rd Chinese Control and Decision Conference, CCDC 2021 - Kunming, China Duration: 22 May 2021 → 24 May 2021

Publication series

Name	Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021

Conference

Conference	33rd Chinese Control and Decision Conference, CCDC 2021
Country/Territory	China
City	Kunming
Period	22/05/21 → 24/05/21

Keywords

Action recognition
Attention mechanism
Long short-term memory network
Residual network

Access to Document

10.1109/CCDC52312.2021.9601577

Cite this

Zhang, Y., & Dong, P. (2021). Video action recognition method based on attention residual network and LSTM. In Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021 (pp. 3611-3616). (Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CCDC52312.2021.9601577

@inproceedings{68dc9c48a5774331adefcee1a657b568,

title = "Video action recognition method based on attention residual network and LSTM",

abstract = "A video action recognition method based on attention residual network and long-term memory network(LSTM) is proposed, which is to solve the problems that the existing human action recognition methods are prone to overfitting, susceptible to interference information, and lack of feature expression ability. In the beginning, the traditional data preprocessing method and sampling method are improved to enhance the generalization ability of the model. Then, a residual network with attention is proposed to improve the feature extraction ability of the network. At length, LSTM is used to recognize video actions. Experimental results on UCF YouTube dataset show that the proposed method can recognize the actions in video more effectively than other similar methods in this field, and the recognition rate reaches 95.45%.",

keywords = "Action recognition, Attention mechanism, Long short-term memory network, Residual network",

author = "Yu Zhang and Pengyue Dong",

note = "Publisher Copyright: {\textcopyright} 2021 IEEE.; 33rd Chinese Control and Decision Conference, CCDC 2021 ; Conference date: 22-05-2021 Through 24-05-2021",

year = "2021",

doi = "10.1109/CCDC52312.2021.9601577",

language = "English",

series = "Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "3611--3616",

booktitle = "Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021",

address = "United States",

}

Zhang, Y & Dong, P 2021, Video action recognition method based on attention residual network and LSTM. in Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021. Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021, Institute of Electrical and Electronics Engineers Inc., pp. 3611-3616, 33rd Chinese Control and Decision Conference, CCDC 2021, Kunming, China, 22/05/21. https://doi.org/10.1109/CCDC52312.2021.9601577

Video action recognition method based on attention residual network and LSTM. / Zhang, Yu; Dong, Pengyue.
Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021. Institute of Electrical and Electronics Engineers Inc., 2021. p. 3611-3616 (Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Video action recognition method based on attention residual network and LSTM

AU - Zhang, Yu

AU - Dong, Pengyue

PY - 2021

Y1 - 2021

N2 - A video action recognition method based on attention residual network and long-term memory network(LSTM) is proposed, which is to solve the problems that the existing human action recognition methods are prone to overfitting, susceptible to interference information, and lack of feature expression ability. In the beginning, the traditional data preprocessing method and sampling method are improved to enhance the generalization ability of the model. Then, a residual network with attention is proposed to improve the feature extraction ability of the network. At length, LSTM is used to recognize video actions. Experimental results on UCF YouTube dataset show that the proposed method can recognize the actions in video more effectively than other similar methods in this field, and the recognition rate reaches 95.45%.

AB - A video action recognition method based on attention residual network and long-term memory network(LSTM) is proposed, which is to solve the problems that the existing human action recognition methods are prone to overfitting, susceptible to interference information, and lack of feature expression ability. In the beginning, the traditional data preprocessing method and sampling method are improved to enhance the generalization ability of the model. Then, a residual network with attention is proposed to improve the feature extraction ability of the network. At length, LSTM is used to recognize video actions. Experimental results on UCF YouTube dataset show that the proposed method can recognize the actions in video more effectively than other similar methods in this field, and the recognition rate reaches 95.45%.

KW - Action recognition

KW - Attention mechanism

KW - Long short-term memory network

KW - Residual network

UR - http://www.scopus.com/inward/record.url?scp=85125193576&partnerID=8YFLogxK

U2 - 10.1109/CCDC52312.2021.9601577

DO - 10.1109/CCDC52312.2021.9601577

M3 - Conference contribution

AN - SCOPUS:85125193576

T3 - Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021

SP - 3611

EP - 3616

BT - Proceedings of the 33rd Chinese Control and Decision Conference, CCDC 2021

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 33rd Chinese Control and Decision Conference, CCDC 2021

Y2 - 22 May 2021 through 24 May 2021

ER -

Video action recognition method based on attention residual network and LSTM

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this