Online Multiperson Tracking With Occlusion Reasoning and Unsupervised Track Motion Model

Niall McLaughlin; Jesus Martinez-del-Rincon; Paul Miller

doi:10.1109/AVSS.2013.6636613

Online Multiperson Tracking With Occlusion Reasoning and Unsupervised Track Motion Model

Niall McLaughlin, Jesus Martinez-del-Rincon, Paul Miller

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

9 Citations (Scopus)

539 Downloads (Pure)

Abstract

We address the problem of multi-target tracking in realistic crowded conditions by introducing a novel dual-stage online tracking algorithm. The problem of data-association between tracks and detections, based on appearance, is often complicated by partial occlusion. In the first stage, we address the issue of occlusion with a novel method of robust data-association, that can be used to compute the appearance similarity between tracks and detections without the need for explicit knowledge of the occluded regions. In the second stage, broken tracks are linked based on motion and appearance, using an online-learned linking model. The online-learned motion-model for track linking uses the confident tracks from the first stage tracker as training examples. The new approach has been tested on the town centre dataset and has performance comparable with the present state-of-the-art

Original language	English
Title of host publication	Proceedings of the IEEE International Conference on Advanced Video and Signal Based Surveillance
Pages	37-42
Number of pages	6
DOIs	https://doi.org/10.1109/AVSS.2013.6636613
Publication status	Published - Aug 2013
Event	The IEEE International Conference on Advanced Video and Signal Based Surveillance 2013 - Krakov, Poland Duration: 27 Aug 2013 → 30 Aug 2013

Conference

Conference	The IEEE International Conference on Advanced Video and Signal Based Surveillance 2013
Country/Territory	Poland
City	Krakov
Period	27/08/2013 → 30/08/2013

Access to Document

10.1109/AVSS.2013.6636613

Online Multiperson TrackingWith Occlusion Reasoning and Unsupervised Track Motion Model
© 2013 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works
Submitted manuscript, 874 KB

R1118ECI: Centre for Secure Information Technologies (CSIT)
McCanny, J. V. (PI), Cowan, C. (CoI), Crookes, D. (CoI), Fusco, V. (CoI), Linton, D. (CoI), Liu, W. (CoI), Miller, P. (CoI), O'Neill, M. (CoI), Scanlon, W. (CoI) & Sezer, S. (CoI)
01/08/2009 → 30/06/2014
Project: Research

Cite this

@inproceedings{ae6aa6e2e8d34e7681acf06211eb25bc,

title = "Online Multiperson Tracking With Occlusion Reasoning and Unsupervised Track Motion Model",

abstract = "We address the problem of multi-target tracking in realistic crowded conditions by introducing a novel dual-stage online tracking algorithm. The problem of data-association between tracks and detections, based on appearance, is often complicated by partial occlusion. In the first stage, we address the issue of occlusion with a novel method of robust data-association, that can be used to compute the appearance similarity between tracks and detections without the need for explicit knowledge of the occluded regions. In the second stage, broken tracks are linked based on motion and appearance, using an online-learned linking model. The online-learned motion-model for track linking uses the confident tracks from the first stage tracker as training examples. The new approach has been tested on the town centre dataset and has performance comparable with the present state-of-the-art",

author = "Niall McLaughlin and Jesus Martinez-del-Rincon and Paul Miller",

year = "2013",

month = aug,

doi = "10.1109/AVSS.2013.6636613",

language = "English",

isbn = "978-1-4799-0703-8",

pages = "37--42",

booktitle = "Proceedings of the IEEE International Conference on Advanced Video and Signal Based Surveillance",

note = "The IEEE International Conference on Advanced Video and Signal Based Surveillance 2013 ; Conference date: 27-08-2013 Through 30-08-2013",

}

McLaughlin, N , Martinez-del-Rincon, J & Miller, P 2013, Online Multiperson Tracking With Occlusion Reasoning and Unsupervised Track Motion Model. in Proceedings of the IEEE International Conference on Advanced Video and Signal Based Surveillance. pp. 37-42, The IEEE International Conference on Advanced Video and Signal Based Surveillance 2013, Krakov, Poland, 27/08/2013. https://doi.org/10.1109/AVSS.2013.6636613

TY - GEN

T1 - Online Multiperson Tracking With Occlusion Reasoning and Unsupervised Track Motion Model

AU - McLaughlin, Niall

AU - Martinez-del-Rincon, Jesus

AU - Miller, Paul

PY - 2013/8

Y1 - 2013/8

N2 - We address the problem of multi-target tracking in realistic crowded conditions by introducing a novel dual-stage online tracking algorithm. The problem of data-association between tracks and detections, based on appearance, is often complicated by partial occlusion. In the first stage, we address the issue of occlusion with a novel method of robust data-association, that can be used to compute the appearance similarity between tracks and detections without the need for explicit knowledge of the occluded regions. In the second stage, broken tracks are linked based on motion and appearance, using an online-learned linking model. The online-learned motion-model for track linking uses the confident tracks from the first stage tracker as training examples. The new approach has been tested on the town centre dataset and has performance comparable with the present state-of-the-art

AB - We address the problem of multi-target tracking in realistic crowded conditions by introducing a novel dual-stage online tracking algorithm. The problem of data-association between tracks and detections, based on appearance, is often complicated by partial occlusion. In the first stage, we address the issue of occlusion with a novel method of robust data-association, that can be used to compute the appearance similarity between tracks and detections without the need for explicit knowledge of the occluded regions. In the second stage, broken tracks are linked based on motion and appearance, using an online-learned linking model. The online-learned motion-model for track linking uses the confident tracks from the first stage tracker as training examples. The new approach has been tested on the town centre dataset and has performance comparable with the present state-of-the-art

U2 - 10.1109/AVSS.2013.6636613

DO - 10.1109/AVSS.2013.6636613

M3 - Conference contribution

SN - 978-1-4799-0703-8

SP - 37

EP - 42

BT - Proceedings of the IEEE International Conference on Advanced Video and Signal Based Surveillance

T2 - The IEEE International Conference on Advanced Video and Signal Based Surveillance 2013

Y2 - 27 August 2013 through 30 August 2013

ER -

Online Multiperson Tracking With Occlusion Reasoning and Unsupervised Track Motion Model

Abstract

Conference

Access to Document

Fingerprint

Projects

R1118ECI: Centre for Secure Information Technologies (CSIT)

Cite this