Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging

Wencheng Duan; Tao Yang; Jing Li; Yanning Zhang

doi:10.1007/978-3-319-46654-5_80

Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging

Wencheng Duan, Tao Yang, Jing Li, Yanning Zhang

School of Computer Science

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

Abstract

Multiple pedestrian tracking is regarded as a challenging work due to difficulties of occlusion, abrupt motion and changes in appearance. In this paper, we propose a multi-layer graph based data association framework to address occlusion problem. Our framework is hierarchical with three association layers and each layer has its corresponding association method. We generate short tracklets and segment some of them into small pieces based on the segmentation condition in the first layer. The segmented tracklets are merged into long tracklets using spatial-temporal information in the second layer. In the last layer, tracklets in neighboring frame-window are merged to form object track mainly by searching the global maximum overlap ratio of the tracklets. Since appearance information is not available in various scenarios, we don’t use any appearance features in our work. We evaluate our algorithm on extensive sequences including two categories and demonstrate superior experimental results.

Original language	English
Title of host publication	Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings
Editors	Shiguang Shan, Zhisheng You, Jie Zhou, Weishi Zheng, Yunhong Wang, Zhenan Sun, Jianjiang Feng, Qijun Zhao
Publisher	Springer Verlag
Pages	728-735
Number of pages	8
ISBN (Print)	9783319466538
DOIs	https://doi.org/10.1007/978-3-319-46654-5_80
State	Published - 2016
Event	11th Chinese Conference on Biometric Recognition, CCBR 2016 - Chengdu, China Duration: 14 Oct 2016 → 16 Oct 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9967 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th Chinese Conference on Biometric Recognition, CCBR 2016
Country/Territory	China
City	Chengdu
Period	14/10/16 → 16/10/16

Keywords

Multi-layer graph based
Multiple pedestrian tracking
Tracklet segmentation and merging

Access to Document

10.1007/978-3-319-46654-5_80

Cite this

Duan, W., Yang, T., Li, J., & Zhang, Y. (2016). Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging. In S. Shan, Z. You, J. Zhou, W. Zheng, Y. Wang, Z. Sun, J. Feng, & Q. Zhao (Eds.), Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings (pp. 728-735). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9967 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46654-5_80

Duan, Wencheng ; Yang, Tao ; Li, Jing et al. / Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging. Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings. editor / Shiguang Shan ; Zhisheng You ; Jie Zhou ; Weishi Zheng ; Yunhong Wang ; Zhenan Sun ; Jianjiang Feng ; Qijun Zhao. Springer Verlag, 2016. pp. 728-735 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{64e29d0010474c51ae6d741083eacfac,

title = "Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging",

abstract = "Multiple pedestrian tracking is regarded as a challenging work due to difficulties of occlusion, abrupt motion and changes in appearance. In this paper, we propose a multi-layer graph based data association framework to address occlusion problem. Our framework is hierarchical with three association layers and each layer has its corresponding association method. We generate short tracklets and segment some of them into small pieces based on the segmentation condition in the first layer. The segmented tracklets are merged into long tracklets using spatial-temporal information in the second layer. In the last layer, tracklets in neighboring frame-window are merged to form object track mainly by searching the global maximum overlap ratio of the tracklets. Since appearance information is not available in various scenarios, we don{\textquoteright}t use any appearance features in our work. We evaluate our algorithm on extensive sequences including two categories and demonstrate superior experimental results.",

keywords = "Multi-layer graph based, Multiple pedestrian tracking, Tracklet segmentation and merging",

author = "Wencheng Duan and Tao Yang and Jing Li and Yanning Zhang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2016.; 11th Chinese Conference on Biometric Recognition, CCBR 2016 ; Conference date: 14-10-2016 Through 16-10-2016",

year = "2016",

doi = "10.1007/978-3-319-46654-5_80",

language = "英语",

isbn = "9783319466538",

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

publisher = "Springer Verlag",

pages = "728--735",

editor = "Shiguang Shan and Zhisheng You and Jie Zhou and Weishi Zheng and Yunhong Wang and Zhenan Sun and Jianjiang Feng and Qijun Zhao",

booktitle = "Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings",

}

Duan, W, Yang, T, Li, J & Zhang, Y 2016, Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging. in S Shan, Z You, J Zhou, W Zheng, Y Wang, Z Sun, J Feng & Q Zhao (eds), Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9967 LNCS, Springer Verlag, pp. 728-735, 11th Chinese Conference on Biometric Recognition, CCBR 2016, Chengdu, China, 14/10/16. https://doi.org/10.1007/978-3-319-46654-5_80

Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging. / Duan, Wencheng; Yang, Tao; Li, Jing et al.
Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings. ed. / Shiguang Shan; Zhisheng You; Jie Zhou; Weishi Zheng; Yunhong Wang; Zhenan Sun; Jianjiang Feng; Qijun Zhao. Springer Verlag, 2016. p. 728-735 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9967 LNCS).

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

TY - GEN

T1 - Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging

AU - Duan, Wencheng

AU - Yang, Tao

AU - Li, Jing

AU - Zhang, Yanning

N1 - Publisher Copyright: © Springer International Publishing AG 2016.

PY - 2016

Y1 - 2016

N2 - Multiple pedestrian tracking is regarded as a challenging work due to difficulties of occlusion, abrupt motion and changes in appearance. In this paper, we propose a multi-layer graph based data association framework to address occlusion problem. Our framework is hierarchical with three association layers and each layer has its corresponding association method. We generate short tracklets and segment some of them into small pieces based on the segmentation condition in the first layer. The segmented tracklets are merged into long tracklets using spatial-temporal information in the second layer. In the last layer, tracklets in neighboring frame-window are merged to form object track mainly by searching the global maximum overlap ratio of the tracklets. Since appearance information is not available in various scenarios, we don’t use any appearance features in our work. We evaluate our algorithm on extensive sequences including two categories and demonstrate superior experimental results.

AB - Multiple pedestrian tracking is regarded as a challenging work due to difficulties of occlusion, abrupt motion and changes in appearance. In this paper, we propose a multi-layer graph based data association framework to address occlusion problem. Our framework is hierarchical with three association layers and each layer has its corresponding association method. We generate short tracklets and segment some of them into small pieces based on the segmentation condition in the first layer. The segmented tracklets are merged into long tracklets using spatial-temporal information in the second layer. In the last layer, tracklets in neighboring frame-window are merged to form object track mainly by searching the global maximum overlap ratio of the tracklets. Since appearance information is not available in various scenarios, we don’t use any appearance features in our work. We evaluate our algorithm on extensive sequences including two categories and demonstrate superior experimental results.

KW - Multi-layer graph based

KW - Multiple pedestrian tracking

KW - Tracklet segmentation and merging

UR - http://www.scopus.com/inward/record.url?scp=84992446851&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-46654-5_80

DO - 10.1007/978-3-319-46654-5_80

M3 - 会议稿件

AN - SCOPUS:84992446851

SN - 9783319466538

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

SP - 728

EP - 735

BT - Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings

A2 - Shan, Shiguang

A2 - You, Zhisheng

A2 - Zhou, Jie

A2 - Zheng, Weishi

A2 - Wang, Yunhong

A2 - Sun, Zhenan

A2 - Feng, Jianjiang

A2 - Zhao, Qijun

PB - Springer Verlag

T2 - 11th Chinese Conference on Biometric Recognition, CCBR 2016

Y2 - 14 October 2016 through 16 October 2016

ER -

Duan W, Yang T, Li J, Zhang Y. Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging. In Shan S, You Z, Zhou J, Zheng W, Wang Y, Sun Z, Feng J, Zhao Q, editors, Biometric Recognition - 11th Chinese Conference, CCBR 2016, Proceedings. Springer Verlag. 2016. p. 728-735. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-46654-5_80

Multiple pedestrian tracking based on multi-layer graph with tracklet segmentation and merging

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this