Mattia Segu (PhD Student)

MSc Mattia Segu

Address: Max-Planck-Institut für Informatik
Saarland Informatics Campus
Campus E1 4
66123 Saarbrücken
Standort: E1 4 - 622
Telefon: +49 681 9325 0
Fax: +49 681 9325 2099
E-mail: Get email via email

Personal Information

Publications

2024

Conference paper

M. Segu, L. Piccinelli, S. Li, L. V. Gool, F. Yu, and B. Schiele

“Walker: Self-supervised Multiple Object Tracking by Walking on Temporal Appearance Graphs,” in Computer Vision -- ECCV 2024, Milan, Italy, 2024.

mehr

BibTeX

@inproceedings{Segu_ECCV24,
TITLE = {Walker: {S}elf-supervised Multiple Object Tracking by Walking on Temporal Appearance Graphs},
AUTHOR = {Segu, Mattia and Piccinelli, Luigi and Li, Siyuan and Gool, Luc Van and Yu, Fisher and Schiele, B.},
LANGUAGE = {eng},
ISBN = {978-3-031-73241-6},
DOI = {10.1007/978-3-031-73242-3_1},
PUBLISHER = {Springer},
YEAR = {2024},
MARGINALMARK = {$\bullet$},
DATE = {2024},
BOOKTITLE = {Computer Vision -- ECCV 2024},
EDITOR = {Leonardis, Ale{\v s} and Ricci, Elisa and Roth, Stefan and Russakovsky, Olga and Sattler, Torsten and Varel, G{\"u}l},
PAGES = {1--18},
SERIES = {Lecture Notes in Compute Science},
VOLUME = {15066},
ADDRESS = {Milan, Italy},
}

Endnote

%0 Conference Proceedings
%A Segu, Mattia
%A Piccinelli, Luigi
%A Li, Siyuan
%A Gool, Luc Van
%A Yu, Fisher
%A Schiele, B.
%+ Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
External Organizations
External Organizations
External Organizations
External Organizations
Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
%T Walker: Self-supervised Multiple Object Tracking by Walking on Temporal Appearance Graphs : 
%G eng
%U http://hdl.handle.net/21.11116/0000-0010-0FCA-B
%R 10.1007/978-3-031-73242-3_1
%D 2024
%B 18th European Conference on Computer Vision
%Z date of event: 2024-09-29 - 2024-10-04
%C Milan, Italy
%B Computer Vision -- ECCV 2024
%E Leonardis, Ale&#353;; Ricci, Elisa; Roth, Stefan; Russakovsky, Olga; Sattler, Torsten; Varel, G&#252;l
%P 1 - 18
%I Springer
%@ 978-3-031-73241-6
%B Lecture Notes in Compute Science
%N 15066
%U https://rdcu.be/dZVCw

Paper

M. Segu, L. Piccinelli, S. Li, Y.-H. Yang, B. Schiele, and L. Van Gool

“Samba: Synchronized Set-of-Sequences Modeling for Multiple Object Tracking,” 2024. [Online]. Available: https://arxiv.org/abs/2410.01806.

mehr

Abstract

Multiple object tracking in complex scenarios - such as coordinated dance
performances, team sports, or dynamic animal groups - presents unique
challenges. In these settings, objects frequently move in coordinated patterns,
occlude each other, and exhibit long-term dependencies in their trajectories.
However, it remains a key open research question on how to model long-range
dependencies within tracklets, interdependencies among tracklets, and the
associated temporal occlusions. To this end, we introduce Samba, a novel
linear-time set-of-sequences model designed to jointly process multiple
tracklets by synchronizing the multiple selective state-spaces used to model
each tracklet. Samba autoregressively predicts the future track query for each
sequence while maintaining synchronized long-term memory representations across
tracklets. By integrating Samba into a tracking-by-propagation framework, we
propose SambaMOTR, the first tracker effectively addressing the aforementioned
issues, including long-range dependencies, tracklet interdependencies, and
temporal occlusions. Additionally, we introduce an effective technique for
dealing with uncertain observations (MaskObs) and an efficient training recipe
to scale SambaMOTR to longer sequences. By modeling long-range dependencies and
interactions among tracked objects, SambaMOTR implicitly learns to track
objects accurately through occlusions without any hand-crafted heuristics. Our
approach significantly surpasses prior state-of-the-art on the DanceTrack, BFT,
and SportsMOT datasets.

BibTeX

@online{Segu_2410.01806,
TITLE = {Samba: Synchronized Set-of-Sequences Modeling for Multiple Object Tracking},
AUTHOR = {Segu, Mattia and Piccinelli, Luigi and Li, Siyuan and Yang, Yung-Hsu and Schiele, Bernt and Van Gool, Luc},
LANGUAGE = {eng},
URL = {https://arxiv.org/abs/2410.01806},
EPRINT = {2410.01806},
EPRINTTYPE = {arXiv},
YEAR = {2024},
MARGINALMARK = {$\bullet$},
ABSTRACT = {Multiple object tracking in complex scenarios -- such as coordinated dance<br>performances, team sports, or dynamic animal groups -- presents unique<br>challenges. In these settings, objects frequently move in coordinated patterns,<br>occlude each other, and exhibit long-term dependencies in their trajectories.<br>However, it remains a key open research question on how to model long-range<br>dependencies within tracklets, interdependencies among tracklets, and the<br>associated temporal occlusions. To this end, we introduce Samba, a novel<br>linear-time set-of-sequences model designed to jointly process multiple<br>tracklets by synchronizing the multiple selective state-spaces used to model<br>each tracklet. Samba autoregressively predicts the future track query for each<br>sequence while maintaining synchronized long-term memory representations across<br>tracklets. By integrating Samba into a tracking-by-propagation framework, we<br>propose SambaMOTR, the first tracker effectively addressing the aforementioned<br>issues, including long-range dependencies, tracklet interdependencies, and<br>temporal occlusions. Additionally, we introduce an effective technique for<br>dealing with uncertain observations (MaskObs) and an efficient training recipe<br>to scale SambaMOTR to longer sequences. By modeling long-range dependencies and<br>interactions among tracked objects, SambaMOTR implicitly learns to track<br>objects accurately through occlusions without any hand-crafted heuristics. Our<br>approach significantly surpasses prior state-of-the-art on the DanceTrack, BFT,<br>and SportsMOT datasets.<br>},
}

Endnote

%0 Report
%A Segu, Mattia
%A Piccinelli, Luigi
%A Li, Siyuan
%A Yang, Yung-Hsu
%A Schiele, Bernt
%A Van Gool, Luc
%+ Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
External Organizations
External Organizations
External Organizations
Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
External Organizations
%T Samba: Synchronized Set-of-Sequences Modeling for Multiple Object
  Tracking : 
%G eng
%U http://hdl.handle.net/21.11116/0000-0010-0FC7-E
%U https://arxiv.org/abs/2410.01806
%D 2024
%X   Multiple object tracking in complex scenarios - such as coordinated dance<br>performances, team sports, or dynamic animal groups - presents unique<br>challenges. In these settings, objects frequently move in coordinated patterns,<br>occlude each other, and exhibit long-term dependencies in their trajectories.<br>However, it remains a key open research question on how to model long-range<br>dependencies within tracklets, interdependencies among tracklets, and the<br>associated temporal occlusions. To this end, we introduce Samba, a novel<br>linear-time set-of-sequences model designed to jointly process multiple<br>tracklets by synchronizing the multiple selective state-spaces used to model<br>each tracklet. Samba autoregressively predicts the future track query for each<br>sequence while maintaining synchronized long-term memory representations across<br>tracklets. By integrating Samba into a tracking-by-propagation framework, we<br>propose SambaMOTR, the first tracker effectively addressing the aforementioned<br>issues, including long-range dependencies, tracklet interdependencies, and<br>temporal occlusions. Additionally, we introduce an effective technique for<br>dealing with uncertain observations (MaskObs) and an efficient training recipe<br>to scale SambaMOTR to longer sequences. By modeling long-range dependencies and<br>interactions among tracked objects, SambaMOTR implicitly learns to track<br>objects accurately through occlusions without any hand-crafted heuristics. Our<br>approach significantly surpasses prior state-of-the-art on the DanceTrack, BFT,<br>and SportsMOT datasets.<br>
%K Computer Science, Computer Vision and Pattern Recognition, cs.CV,Computer Science, Artificial Intelligence, cs.AI

2023

Conference paper

Q. Fan, M. Segu, Y.-W. Tai, F. Yu, C.-K. Tang, B. Schiele, and D. Dai

“Towards Robust Object Detection Invariant to Real-World Domain Shifts,” in Eleventh International Conference on Learning Representations (ICLR 2023), Kigali, Rwanda, 2023.

mehr

BibTeX

@inproceedings{Fan_ICLR23,
TITLE = {Towards Robust Object Detection Invariant to Real-World Domain Shifts},
AUTHOR = {Fan, Qi and Segu, Mattia and Tai, Yu-Wing and Yu, Fisher and Tang, Chi-Keung and Schiele, Bernt and Dai, Dengxin},
LANGUAGE = {eng},
URL = {https://openreview.net/group?id=ICLR.cc/2023/Conference#poster},
PUBLISHER = {OpenReview.net},
YEAR = {2023},
MARGINALMARK = {$\bullet$},
BOOKTITLE = {Eleventh International Conference on Learning Representations (ICLR 2023)},
ADDRESS = {Kigali, Rwanda},
}

Endnote

%0 Conference Proceedings
%A Fan, Qi
%A Segu, Mattia
%A Tai, Yu-Wing
%A Yu, Fisher
%A Tang, Chi-Keung
%A Schiele, Bernt
%A Dai, Dengxin
%+ Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
External Organizations
External Organizations
External Organizations
Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
%T Towards Robust Object Detection Invariant to Real-World Domain Shifts : 
%G eng
%U http://hdl.handle.net/21.11116/0000-0010-45CA-D
%D 2023
%B Eleventh International Conference on Learning Representations
%Z date of event: 2023-05-01 - 2023-05-05
%C Kigali, Rwanda
%B Eleventh International Conference on Learning Representations
%I OpenReview.net
%U https://openreview.net/group?id=ICLR.cc/2023/Conference#poster
%U https://openreview.net/forum?id=vqSyt8D3ny

Conference paper

M. Segu, B. Schiele, and F. Yu

“DARTH: Holistic Test-time Adaptation for Multiple Object Tracking,” in IEEE/CVF International Conference on Computer Vision (ICCV 2023), Paris, France, 2023.

mehr

BibTeX

@inproceedings{Segu_ICCV23,
TITLE = {{DARTH}: {H}olistic Test-time Adaptation for Multiple Object Tracking},
AUTHOR = {Segu, Mattia and Schiele, Bernt and Yu, Fisher},
LANGUAGE = {eng},
ISBN = {979-8-3503-0718-4},
DOI = {10.1109/ICCV51070.2023.00891},
PUBLISHER = {IEEE},
YEAR = {2023},
MARGINALMARK = {$\bullet$},
DATE = {2023},
BOOKTITLE = {IEEE/CVF International Conference on Computer Vision (ICCV 2023)},
PAGES = {9683--9693},
ADDRESS = {Paris, France},
}

Endnote

%0 Conference Proceedings
%A Segu, Mattia
%A Schiele, Bernt
%A Yu, Fisher
%+ Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
Computer Vision and Machine Learning, MPI for Informatics, Max Planck Society
External Organizations
%T DARTH: Holistic Test-time Adaptation for Multiple Object Tracking : 
%G eng
%U http://hdl.handle.net/21.11116/0000-000D-D1AB-6
%R 10.1109/ICCV51070.2023.00891
%D 2023
%B IEEE/CVF International Conference on Computer Vision
%Z date of event: 2023-10-02 - 2023-10-06
%C Paris, France
%B IEEE/CVF International Conference on Computer Vision 
%P 9683 - 9693
%I IEEE
%@ 979-8-3503-0718-4