Generalizable Online 3D Pedestrian Tracking with Multiple Cameras

Victor Lyra; Isabella de Andrade; João Paulo Lima; Rafael Roberto; Lucas Figueiredo; João Marcelo Teixeira; Diego Thomas; Hideaki Uchiyama; Veronica Teichrieb

doi:10.5220/0010842800003124

Generalizable Online 3D Pedestrian Tracking with Multiple Cameras

Victor Lyra, Isabella de Andrade, João Paulo Lima, Rafael Roberto, Lucas Figueiredo, João Marcelo Teixeira, Diego Thomas, Hideaki Uchiyama, Veronica Teichrieb

Department of Mechanical Engineering

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

4 Citations (SciVal)

Abstract

3D pedestrian tracking using multiple cameras is still a challenging task with many applications such as surveillance, behavioral analysis, statistical analysis, and more. Many of the existing tracking solutions involve training the algorithms on the target environment, which requires extensive time and effort. We propose an online 3D pedestrian tracking method for multi-camera environments based on a generalizable detection solution that does not require training with data of the target scene. We establish temporal relationships between people detected in different frames by using a combination of graph matching algorithm and Kalman filter. Our proposed method obtained a MOTA and MOTP of 77.1% and 96.4%, respectively on the test split of the public WILDTRACK dataset. Such results correspond to an improvement of approximately 3.4% and 22.2%, respectively, compared to the best existing online technique. Our experiments also demonstrate the advantages of using appearance information to improve the tracking performance.

Original language	English
Title of host publication	Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022)
Pages	820-827
Number of pages	8
Volume	5
ISBN (Electronic)	9789897585555
DOIs	https://doi.org/10.5220/0010842800003124
Publication status	Published - 8 Feb 2022
Event	17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2022 - Virtual, Online Duration: 6 Feb 2022 → 8 Feb 2022

Publication series

Name	Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications
Publisher	Science and Technology Publications, Lda
ISSN (Print)	2184-5921

Conference

Conference	17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2022
City	Virtual, Online
Period	6/02/22 → 8/02/22

Keywords

Detection
Multiple Cameras
Pedestrians
Tracking

ASJC Scopus subject areas

Computer Graphics and Computer-Aided Design
Computer Vision and Pattern Recognition
Human-Computer Interaction

Access to Document

10.5220/0010842800003124Licence: CC BY-NC-ND

Cite this

Lyra, V., de Andrade, I., Lima, J. P., Roberto, R., Figueiredo, L., Teixeira, J. M., Thomas, D., Uchiyama, H., & Teichrieb, V. (2022). Generalizable Online 3D Pedestrian Tracking with Multiple Cameras. In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) (Vol. 5, pp. 820-827). (Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications). https://doi.org/10.5220/0010842800003124

Generalizable Online 3D Pedestrian Tracking with Multiple Cameras. / Lyra, Victor; de Andrade, Isabella; Lima, João Paulo et al.
Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) . Vol. 5 2022. p. 820-827 (Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications).

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

Lyra, V, de Andrade, I, Lima, JP, Roberto, R, Figueiredo, L, Teixeira, JM, Thomas, D, Uchiyama, H & Teichrieb, V 2022, Generalizable Online 3D Pedestrian Tracking with Multiple Cameras. in Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) . vol. 5, Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, pp. 820-827, 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2022, Virtual, Online, 6/02/22. https://doi.org/10.5220/0010842800003124

Lyra V, de Andrade I, Lima JP, Roberto R, Figueiredo L, Teixeira JM et al. Generalizable Online 3D Pedestrian Tracking with Multiple Cameras. In Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) . Vol. 5. 2022. p. 820-827. (Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications). doi: 10.5220/0010842800003124

Lyra, Victor ; de Andrade, Isabella ; Lima, João Paulo et al. / Generalizable Online 3D Pedestrian Tracking with Multiple Cameras. Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022) . Vol. 5 2022. pp. 820-827 (Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications).

@inproceedings{19b9e9955d78460ebb8c41221260b7bf,

title = "Generalizable Online 3D Pedestrian Tracking with Multiple Cameras",

abstract = "3D pedestrian tracking using multiple cameras is still a challenging task with many applications such as surveillance, behavioral analysis, statistical analysis, and more. Many of the existing tracking solutions involve training the algorithms on the target environment, which requires extensive time and effort. We propose an online 3D pedestrian tracking method for multi-camera environments based on a generalizable detection solution that does not require training with data of the target scene. We establish temporal relationships between people detected in different frames by using a combination of graph matching algorithm and Kalman filter. Our proposed method obtained a MOTA and MOTP of 77.1\% and 96.4\%, respectively on the test split of the public WILDTRACK dataset. Such results correspond to an improvement of approximately 3.4\% and 22.2\%, respectively, compared to the best existing online technique. Our experiments also demonstrate the advantages of using appearance information to improve the tracking performance.",

keywords = "Detection, Multiple Cameras, Pedestrians, Tracking",

author = "Victor Lyra and \{de Andrade\}, Isabella and Lima, \{Jo{\~a}o Paulo\} and Rafael Roberto and Lucas Figueiredo and Teixeira, \{Jo{\~a}o Marcelo\} and Diego Thomas and Hideaki Uchiyama and Veronica Teichrieb",

year = "2022",

month = feb,

day = "8",

doi = "10.5220/0010842800003124",

language = "English",

volume = "5",

series = "Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications",

publisher = "Science and Technology Publications, Lda",

pages = "820--827",

booktitle = "Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022)",

note = "17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2022 ; Conference date: 06-02-2022 Through 08-02-2022",

}

TY - GEN

T1 - Generalizable Online 3D Pedestrian Tracking with Multiple Cameras

AU - Lyra, Victor

AU - de Andrade, Isabella

AU - Lima, João Paulo

AU - Roberto, Rafael

AU - Figueiredo, Lucas

AU - Teixeira, João Marcelo

AU - Thomas, Diego

AU - Uchiyama, Hideaki

AU - Teichrieb, Veronica

PY - 2022/2/8

Y1 - 2022/2/8

N2 - 3D pedestrian tracking using multiple cameras is still a challenging task with many applications such as surveillance, behavioral analysis, statistical analysis, and more. Many of the existing tracking solutions involve training the algorithms on the target environment, which requires extensive time and effort. We propose an online 3D pedestrian tracking method for multi-camera environments based on a generalizable detection solution that does not require training with data of the target scene. We establish temporal relationships between people detected in different frames by using a combination of graph matching algorithm and Kalman filter. Our proposed method obtained a MOTA and MOTP of 77.1% and 96.4%, respectively on the test split of the public WILDTRACK dataset. Such results correspond to an improvement of approximately 3.4% and 22.2%, respectively, compared to the best existing online technique. Our experiments also demonstrate the advantages of using appearance information to improve the tracking performance.

AB - 3D pedestrian tracking using multiple cameras is still a challenging task with many applications such as surveillance, behavioral analysis, statistical analysis, and more. Many of the existing tracking solutions involve training the algorithms on the target environment, which requires extensive time and effort. We propose an online 3D pedestrian tracking method for multi-camera environments based on a generalizable detection solution that does not require training with data of the target scene. We establish temporal relationships between people detected in different frames by using a combination of graph matching algorithm and Kalman filter. Our proposed method obtained a MOTA and MOTP of 77.1% and 96.4%, respectively on the test split of the public WILDTRACK dataset. Such results correspond to an improvement of approximately 3.4% and 22.2%, respectively, compared to the best existing online technique. Our experiments also demonstrate the advantages of using appearance information to improve the tracking performance.

KW - Detection

KW - Multiple Cameras

KW - Pedestrians

KW - Tracking

UR - http://www.scopus.com/inward/record.url?scp=85169418833&partnerID=8YFLogxK

U2 - 10.5220/0010842800003124

DO - 10.5220/0010842800003124

M3 - Chapter in a published conference proceeding

AN - SCOPUS:85169418833

VL - 5

T3 - Proceedings of the International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications

SP - 820

EP - 827

BT - Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2022)

T2 - 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, VISIGRAPP 2022

Y2 - 6 February 2022 through 8 February 2022

ER -

Generalizable Online 3D Pedestrian Tracking with Multiple Cameras

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this