Relaxed Pairwise Learned Metric for Person Re-identification

Martin Hirzer, Peter Roth, Martin Köstinger, Horst Bischof

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

Matching persons across non-overlapping cameras is a rather challenging task. Thus, successful methods often build on complex feature representations or sophisticated learners. A recent trend to tackle this problem is to use metric learning to find a suitable space for matching samples from different cameras. However, most of these approaches ignore the transition from one camera to the other. In this paper, we propose to learn a metric from pairs of samples from different cameras. In this way, even less sophisticated features describing color and texture information are sufficient for finally getting state-of-the-art classification results. Moreover, once the metric has been learned, only linear projections are necessary at search time, where a simple nearest neighbor classification is performed. The approach is demonstrated on three publicly available datasets of different complexity, where it can be seen that state-of-the-art results can be obtained at much lower computational costs.
LanguageEnglish
Title of host publicationProceedings of the European Conference on Computer Vision (ECCV)
Publisher.
Pages780-793
DOIs
StatusPublished - 2012
EventEuropean Conference on Computer Vision - Florenz, Italy
Duration: 7 Oct 201213 Oct 2012

Conference

ConferenceEuropean Conference on Computer Vision
Abbreviated titleECCV
CountryItaly
CityFlorenz
Period7/10/1213/10/12

Fingerprint

Cameras
Textures
Color
Costs

Fields of Expertise

  • Information, Communication & Computing

Cite this

Hirzer, M., Roth, P., Köstinger, M., & Bischof, H. (2012). Relaxed Pairwise Learned Metric for Person Re-identification. In Proceedings of the European Conference on Computer Vision (ECCV) (pp. 780-793). .. https://doi.org/10.1007/978-3-642-33783-3_56

Relaxed Pairwise Learned Metric for Person Re-identification. / Hirzer, Martin; Roth, Peter; Köstinger, Martin; Bischof, Horst.

Proceedings of the European Conference on Computer Vision (ECCV). ., 2012. p. 780-793.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Hirzer, M, Roth, P, Köstinger, M & Bischof, H 2012, Relaxed Pairwise Learned Metric for Person Re-identification. in Proceedings of the European Conference on Computer Vision (ECCV). ., pp. 780-793, European Conference on Computer Vision, Florenz, Italy, 7/10/12. https://doi.org/10.1007/978-3-642-33783-3_56
Hirzer M, Roth P, Köstinger M, Bischof H. Relaxed Pairwise Learned Metric for Person Re-identification. In Proceedings of the European Conference on Computer Vision (ECCV). . 2012. p. 780-793 https://doi.org/10.1007/978-3-642-33783-3_56
Hirzer, Martin ; Roth, Peter ; Köstinger, Martin ; Bischof, Horst. / Relaxed Pairwise Learned Metric for Person Re-identification. Proceedings of the European Conference on Computer Vision (ECCV). ., 2012. pp. 780-793
@inproceedings{5c5cd0af64424df893f92de13b04de33,
title = "Relaxed Pairwise Learned Metric for Person Re-identification",
abstract = "Matching persons across non-overlapping cameras is a rather challenging task. Thus, successful methods often build on complex feature representations or sophisticated learners. A recent trend to tackle this problem is to use metric learning to find a suitable space for matching samples from different cameras. However, most of these approaches ignore the transition from one camera to the other. In this paper, we propose to learn a metric from pairs of samples from different cameras. In this way, even less sophisticated features describing color and texture information are sufficient for finally getting state-of-the-art classification results. Moreover, once the metric has been learned, only linear projections are necessary at search time, where a simple nearest neighbor classification is performed. The approach is demonstrated on three publicly available datasets of different complexity, where it can be seen that state-of-the-art results can be obtained at much lower computational costs.",
author = "Martin Hirzer and Peter Roth and Martin K{\"o}stinger and Horst Bischof",
year = "2012",
doi = "10.1007/978-3-642-33783-3_56",
language = "English",
pages = "780--793",
booktitle = "Proceedings of the European Conference on Computer Vision (ECCV)",
publisher = ".",

}

TY - GEN

T1 - Relaxed Pairwise Learned Metric for Person Re-identification

AU - Hirzer, Martin

AU - Roth, Peter

AU - Köstinger, Martin

AU - Bischof, Horst

PY - 2012

Y1 - 2012

N2 - Matching persons across non-overlapping cameras is a rather challenging task. Thus, successful methods often build on complex feature representations or sophisticated learners. A recent trend to tackle this problem is to use metric learning to find a suitable space for matching samples from different cameras. However, most of these approaches ignore the transition from one camera to the other. In this paper, we propose to learn a metric from pairs of samples from different cameras. In this way, even less sophisticated features describing color and texture information are sufficient for finally getting state-of-the-art classification results. Moreover, once the metric has been learned, only linear projections are necessary at search time, where a simple nearest neighbor classification is performed. The approach is demonstrated on three publicly available datasets of different complexity, where it can be seen that state-of-the-art results can be obtained at much lower computational costs.

AB - Matching persons across non-overlapping cameras is a rather challenging task. Thus, successful methods often build on complex feature representations or sophisticated learners. A recent trend to tackle this problem is to use metric learning to find a suitable space for matching samples from different cameras. However, most of these approaches ignore the transition from one camera to the other. In this paper, we propose to learn a metric from pairs of samples from different cameras. In this way, even less sophisticated features describing color and texture information are sufficient for finally getting state-of-the-art classification results. Moreover, once the metric has been learned, only linear projections are necessary at search time, where a simple nearest neighbor classification is performed. The approach is demonstrated on three publicly available datasets of different complexity, where it can be seen that state-of-the-art results can be obtained at much lower computational costs.

U2 - 10.1007/978-3-642-33783-3_56

DO - 10.1007/978-3-642-33783-3_56

M3 - Conference contribution

SP - 780

EP - 793

BT - Proceedings of the European Conference on Computer Vision (ECCV)

PB - .

ER -