Improving Global Multi-target Tracking with Local Updates

Anton Milan, Rikke Gade, Anthony Dick, Thomas B. Moeslund, Ian Reid

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

7 Citationer (Scopus)
313 Downloads (Pure)

Resumé

We propose a scheme to explicitly detect and resolve ambiguous situations in multiple target tracking. During periods of uncertainty, our method applies multiple local single target trackers to hypothesise short term tracks. These tracks are combined with the tracks obtained by a global multi-target tracker, if they result in a reduction in the global cost function. Since tracking failures typically arise when targets become occluded, we propose a local data association scheme to maintain the target identities in these situations. We demonstrate a reduction of up to 50% in the global cost function, which in turn leads to superior performance on several challenging benchmark sequences. Additionally, we show tracking results in sports videos where poor video quality and frequent and severe occlusions between multiple players pose difficulties for
state-of-the-art trackers.
OriginalsprogEngelsk
TitelComputer Vision - ECCV 2014 Workshops : Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part III
RedaktørerLourdes Agapito, Michael M. Bronstein, Carsten Rother
Antal sider17
ForlagSpringer
Publikationsdato2015
Udgave3
Sider174-190
ISBN (Trykt)978-3-319-16198-3
ISBN (Elektronisk)978-3-319-16199-0
DOI
StatusUdgivet - 2015
BegivenhedECCV Workshop: Visual Surveillance and Re-Identification - Zürich, Schweiz
Varighed: 12 sep. 2014 → …

Workshop

WorkshopECCV Workshop
LandSchweiz
ByZürich
Periode12/09/2014 → …
NavnLecture Notes in Computer Science
Nummer8927
ISSN0302-9743

Fingerprint

Target tracking
Cost functions
Sports
Association reactions
Uncertainty

Citer dette

Milan, A., Gade, R., Dick, A., Moeslund, T. B., & Reid, I. (2015). Improving Global Multi-target Tracking with Local Updates. I L. Agapito, M. M. Bronstein, & C. Rother (red.), Computer Vision - ECCV 2014 Workshops : Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part III (3 udg., s. 174-190). Springer. Lecture Notes in Computer Science, Nr. 8927 https://doi.org/10.1007/978-3-319-16199-0_13
Milan, Anton ; Gade, Rikke ; Dick, Anthony ; Moeslund, Thomas B. ; Reid, Ian. / Improving Global Multi-target Tracking with Local Updates. Computer Vision - ECCV 2014 Workshops : Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part III. red. / Lourdes Agapito ; Michael M. Bronstein ; Carsten Rother. 3. udg. Springer, 2015. s. 174-190 (Lecture Notes in Computer Science; Nr. 8927).
@inproceedings{c2791167033d4cb28cb70d0145ef7932,
title = "Improving Global Multi-target Tracking with Local Updates",
abstract = "We propose a scheme to explicitly detect and resolve ambiguous situations in multiple target tracking. During periods of uncertainty, our method applies multiple local single target trackers to hypothesise short term tracks. These tracks are combined with the tracks obtained by a global multi-target tracker, if they result in a reduction in the global cost function. Since tracking failures typically arise when targets become occluded, we propose a local data association scheme to maintain the target identities in these situations. We demonstrate a reduction of up to 50{\%} in the global cost function, which in turn leads to superior per- formance on several challenging benchmark sequences. Additionally, we show tracking results in sports videos where poor video quality and fre- quent and severe occlusions between multiple players pose difficulties for state-of-the-art trackers.",
author = "Anton Milan and Rikke Gade and Anthony Dick and Moeslund, {Thomas B.} and Ian Reid",
year = "2015",
doi = "10.1007/978-3-319-16199-0_13",
language = "English",
isbn = "978-3-319-16198-3",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
number = "8927",
pages = "174--190",
editor = "Lourdes Agapito and Bronstein, {Michael M.} and Carsten Rother",
booktitle = "Computer Vision - ECCV 2014 Workshops",
address = "Germany",
edition = "3",

}

Milan, A, Gade, R, Dick, A, Moeslund, TB & Reid, I 2015, Improving Global Multi-target Tracking with Local Updates. i L Agapito, MM Bronstein & C Rother (red), Computer Vision - ECCV 2014 Workshops : Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part III. 3 udg, Springer, Lecture Notes in Computer Science, nr. 8927, s. 174-190, ECCV Workshop, Zürich, Schweiz, 12/09/2014. https://doi.org/10.1007/978-3-319-16199-0_13

Improving Global Multi-target Tracking with Local Updates. / Milan, Anton; Gade, Rikke; Dick, Anthony; Moeslund, Thomas B.; Reid, Ian.

Computer Vision - ECCV 2014 Workshops : Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part III. red. / Lourdes Agapito; Michael M. Bronstein; Carsten Rother. 3. udg. Springer, 2015. s. 174-190 (Lecture Notes in Computer Science; Nr. 8927).

Publikation: Bidrag til bog/antologi/rapport/konference proceedingKonferenceartikel i proceedingForskningpeer review

TY - GEN

T1 - Improving Global Multi-target Tracking with Local Updates

AU - Milan, Anton

AU - Gade, Rikke

AU - Dick, Anthony

AU - Moeslund, Thomas B.

AU - Reid, Ian

PY - 2015

Y1 - 2015

N2 - We propose a scheme to explicitly detect and resolve ambiguous situations in multiple target tracking. During periods of uncertainty, our method applies multiple local single target trackers to hypothesise short term tracks. These tracks are combined with the tracks obtained by a global multi-target tracker, if they result in a reduction in the global cost function. Since tracking failures typically arise when targets become occluded, we propose a local data association scheme to maintain the target identities in these situations. We demonstrate a reduction of up to 50% in the global cost function, which in turn leads to superior per- formance on several challenging benchmark sequences. Additionally, we show tracking results in sports videos where poor video quality and fre- quent and severe occlusions between multiple players pose difficulties for state-of-the-art trackers.

AB - We propose a scheme to explicitly detect and resolve ambiguous situations in multiple target tracking. During periods of uncertainty, our method applies multiple local single target trackers to hypothesise short term tracks. These tracks are combined with the tracks obtained by a global multi-target tracker, if they result in a reduction in the global cost function. Since tracking failures typically arise when targets become occluded, we propose a local data association scheme to maintain the target identities in these situations. We demonstrate a reduction of up to 50% in the global cost function, which in turn leads to superior per- formance on several challenging benchmark sequences. Additionally, we show tracking results in sports videos where poor video quality and fre- quent and severe occlusions between multiple players pose difficulties for state-of-the-art trackers.

U2 - 10.1007/978-3-319-16199-0_13

DO - 10.1007/978-3-319-16199-0_13

M3 - Article in proceeding

SN - 978-3-319-16198-3

T3 - Lecture Notes in Computer Science

SP - 174

EP - 190

BT - Computer Vision - ECCV 2014 Workshops

A2 - Agapito, Lourdes

A2 - Bronstein, Michael M.

A2 - Rother, Carsten

PB - Springer

ER -

Milan A, Gade R, Dick A, Moeslund TB, Reid I. Improving Global Multi-target Tracking with Local Updates. I Agapito L, Bronstein MM, Rother C, red., Computer Vision - ECCV 2014 Workshops : Zurich, Switzerland, September 6-7 and 12, 2014, Proceedings, Part III. 3 udg. Springer. 2015. s. 174-190. (Lecture Notes in Computer Science; Nr. 8927). https://doi.org/10.1007/978-3-319-16199-0_13