Improving gait recognition with 3D pose estimation

Weizhi An; Rijun Liao; Shiqi Yu; Yongzhen Huang; Pong Chi YUEN

doi:10.1007/978-3-319-97909-0_15

Improving gait recognition with 3D pose estimation

Weizhi An, Rijun Liao, Shiqi Yu^*, Yongzhen Huang, Pong Chi YUEN

^*Corresponding author for this work

Department of Computer Science

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

39 Citations (Scopus)

Abstract

Gait is a kind of attractive biometric feature for human identification in recent decades. The view, clothing, carrying and other variations are always the challenges for gait recognition. One of the possible solutions is the model based methods. In this paper, 3D pose is estimated from 2D images are used as the feature for gait recognition. So gait can be described by the motion of human body joints. Besides, the 3D pose has better capacity for view variation than the 2D pose. Experimental results also prove that in the paper. To improve the recognition rates, LSTM and CNNs are employed to extract temporal and spatial features. Compared with other model-based methods, the proposed one has achieved much better performance and is comparable with appearance-based ones. The experimental results show the proposed 3D pose based method has unique advantages in large view variation. It will have great potential with the development of pose estimation in future.

Original language	English
Title of host publication	Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings
Editors	Zhenan Sun, Shiguang Shan, Zhenhong Jia, Kurban Ubul, Jie Zhou, Jianjiang Feng, Zhenhua Guo, Yunhong Wang
Publisher	Springer Verlag
Pages	137-147
Number of pages	11
ISBN (Print)	9783319979083
DOIs	https://doi.org/10.1007/978-3-319-97909-0_15
Publication status	Published - 2018
Event	13th Chinese Conference on Biometric Recognition, CCBR 2018 - Urumchi, China Duration: 11 Aug 2018 → 12 Aug 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10996 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	13th Chinese Conference on Biometric Recognition, CCBR 2018
Country/Territory	China
City	Urumchi
Period	11/08/18 → 12/08/18

Scopus Subject Areas

Theoretical Computer Science
Computer Science(all)

User-Defined Keywords

3D pose
CNNs
Gait recognition
LSTM

Access to Document

10.1007/978-3-319-97909-0_15

Cite this

An, W., Liao, R., Yu, S., Huang, Y., & YUEN, P. C. (2018). Improving gait recognition with 3D pose estimation. In Z. Sun, S. Shan, Z. Jia, K. Ubul, J. Zhou, J. Feng, Z. Guo, & Y. Wang (Eds.), Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings (pp. 137-147). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10996 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-97909-0_15

An, Weizhi ; Liao, Rijun ; Yu, Shiqi et al. / Improving gait recognition with 3D pose estimation. Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings. editor / Zhenan Sun ; Shiguang Shan ; Zhenhong Jia ; Kurban Ubul ; Jie Zhou ; Jianjiang Feng ; Zhenhua Guo ; Yunhong Wang. Springer Verlag, 2018. pp. 137-147 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{516fb8146abe4efcb1ff37367f1136a1,

title = "Improving gait recognition with 3D pose estimation",

abstract = "Gait is a kind of attractive biometric feature for human identification in recent decades. The view, clothing, carrying and other variations are always the challenges for gait recognition. One of the possible solutions is the model based methods. In this paper, 3D pose is estimated from 2D images are used as the feature for gait recognition. So gait can be described by the motion of human body joints. Besides, the 3D pose has better capacity for view variation than the 2D pose. Experimental results also prove that in the paper. To improve the recognition rates, LSTM and CNNs are employed to extract temporal and spatial features. Compared with other model-based methods, the proposed one has achieved much better performance and is comparable with appearance-based ones. The experimental results show the proposed 3D pose based method has unique advantages in large view variation. It will have great potential with the development of pose estimation in future.",

keywords = "3D pose, CNNs, Gait recognition, LSTM",

author = "Weizhi An and Rijun Liao and Shiqi Yu and Yongzhen Huang and YUEN, {Pong Chi}",

note = "Funding Information: Acknowledgment. The work is supported by the strategic new and future industrial development fund of Shenzhen (Grant No. 20170504160426188).; 13th Chinese Conference on Biometric Recognition, CCBR 2018 ; Conference date: 11-08-2018 Through 12-08-2018",

year = "2018",

doi = "10.1007/978-3-319-97909-0_15",

language = "English",

isbn = "9783319979083",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "137--147",

editor = "Zhenan Sun and Shiguang Shan and Zhenhong Jia and Kurban Ubul and Jie Zhou and Jianjiang Feng and Zhenhua Guo and Yunhong Wang",

booktitle = "Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings",

address = "Germany",

}

An, W, Liao, R, Yu, S, Huang, Y & YUEN, PC 2018, Improving gait recognition with 3D pose estimation. in Z Sun, S Shan, Z Jia, K Ubul, J Zhou, J Feng, Z Guo & Y Wang (eds), Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10996 LNCS, Springer Verlag, pp. 137-147, 13th Chinese Conference on Biometric Recognition, CCBR 2018, Urumchi, China, 11/08/18. https://doi.org/10.1007/978-3-319-97909-0_15

Improving gait recognition with 3D pose estimation. / An, Weizhi; Liao, Rijun; Yu, Shiqi et al.
Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings. ed. / Zhenan Sun; Shiguang Shan; Zhenhong Jia; Kurban Ubul; Jie Zhou; Jianjiang Feng; Zhenhua Guo; Yunhong Wang. Springer Verlag, 2018. p. 137-147 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10996 LNCS).

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - Improving gait recognition with 3D pose estimation

AU - An, Weizhi

AU - Liao, Rijun

AU - Yu, Shiqi

AU - Huang, Yongzhen

AU - YUEN, Pong Chi

N1 - Funding Information: Acknowledgment. The work is supported by the strategic new and future industrial development fund of Shenzhen (Grant No. 20170504160426188).

PY - 2018

Y1 - 2018

N2 - Gait is a kind of attractive biometric feature for human identification in recent decades. The view, clothing, carrying and other variations are always the challenges for gait recognition. One of the possible solutions is the model based methods. In this paper, 3D pose is estimated from 2D images are used as the feature for gait recognition. So gait can be described by the motion of human body joints. Besides, the 3D pose has better capacity for view variation than the 2D pose. Experimental results also prove that in the paper. To improve the recognition rates, LSTM and CNNs are employed to extract temporal and spatial features. Compared with other model-based methods, the proposed one has achieved much better performance and is comparable with appearance-based ones. The experimental results show the proposed 3D pose based method has unique advantages in large view variation. It will have great potential with the development of pose estimation in future.

AB - Gait is a kind of attractive biometric feature for human identification in recent decades. The view, clothing, carrying and other variations are always the challenges for gait recognition. One of the possible solutions is the model based methods. In this paper, 3D pose is estimated from 2D images are used as the feature for gait recognition. So gait can be described by the motion of human body joints. Besides, the 3D pose has better capacity for view variation than the 2D pose. Experimental results also prove that in the paper. To improve the recognition rates, LSTM and CNNs are employed to extract temporal and spatial features. Compared with other model-based methods, the proposed one has achieved much better performance and is comparable with appearance-based ones. The experimental results show the proposed 3D pose based method has unique advantages in large view variation. It will have great potential with the development of pose estimation in future.

KW - 3D pose

KW - CNNs

KW - Gait recognition

KW - LSTM

UR - http://www.scopus.com/inward/record.url?scp=85051963118&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-97909-0_15

DO - 10.1007/978-3-319-97909-0_15

M3 - Conference proceeding

AN - SCOPUS:85051963118

SN - 9783319979083

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 137

EP - 147

BT - Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings

A2 - Sun, Zhenan

A2 - Shan, Shiguang

A2 - Jia, Zhenhong

A2 - Ubul, Kurban

A2 - Zhou, Jie

A2 - Feng, Jianjiang

A2 - Guo, Zhenhua

A2 - Wang, Yunhong

PB - Springer Verlag

T2 - 13th Chinese Conference on Biometric Recognition, CCBR 2018

Y2 - 11 August 2018 through 12 August 2018

ER -

An W, Liao R, Yu S, Huang Y, YUEN PC. Improving gait recognition with 3D pose estimation. In Sun Z, Shan S, Jia Z, Ubul K, Zhou J, Feng J, Guo Z, Wang Y, editors, Biometric Recognition - 13th Chinese Conference, CCBR 2018, Proceedings. Springer Verlag. 2018. p. 137-147. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-97909-0_15

Improving gait recognition with 3D pose estimation

Abstract

Publication series

Conference

Scopus Subject Areas

User-Defined Keywords

Access to Document

Other files and links

Fingerprint

Cite this