Publication detail

Vehicle Re-Identification for Automatic Video Traffic Surveillance

ZAPLETAL, D. HEROUT, A.

Original Title

Vehicle Re-Identification for Automatic Video Traffic Surveillance

English Title

Vehicle Re-Identification for Automatic Video Traffic Surveillance

Type

conference paper

Language

en

Original Abstract

This paper proposes an approach to the vehicle re-identification problem in a multiple camera system.  We focused on the re-identification itself assuming that the vehicle detection problem is already solved including extraction of a full-fledged 3D bounding box. The re-identification problem is solved by using color histograms and histograms of oriented gradients by a linear regressor.  The features are used in separate models in order to get the best results in the shortest CPU computation time. The proposed method works with a high accuracy (60% true positives retrieved with 10% false positive rate on a challenging subset of the test data) in 85 milliseconds of the CPU (Core i7) computation time per one vehicle re-identification assuming the fullHD resolution video input. The applications of this work include finding important parameters such as travel time, traffic flow, or traffic information in a distributed traffic surveillance and monitoring system.

English abstract

This paper proposes an approach to the vehicle re-identification problem in a multiple camera system.  We focused on the re-identification itself assuming that the vehicle detection problem is already solved including extraction of a full-fledged 3D bounding box. The re-identification problem is solved by using color histograms and histograms of oriented gradients by a linear regressor.  The features are used in separate models in order to get the best results in the shortest CPU computation time. The proposed method works with a high accuracy (60% true positives retrieved with 10% false positive rate on a challenging subset of the test data) in 85 milliseconds of the CPU (Core i7) computation time per one vehicle re-identification assuming the fullHD resolution video input. The applications of this work include finding important parameters such as travel time, traffic flow, or traffic information in a distributed traffic surveillance and monitoring system.

Keywords

vehicle re-identification, traffic monitoring, automatic traffic surveillance

Released

30.06.2016

Publisher

IEEE Computer Society

Location

Las Vegas

ISBN

978-0-7695-4989-7

Book

International Workshop on Automatic Traffic Surveillance (CVPR 2016)

Edition

NEUVEDEN

Edition number

NEUVEDEN

Pages from

1568

Pages to

1574

Pages count

7

Documents

BibTex 


@inproceedings{BUT130978,
  author="Dominik {Zapletal} and Adam {Herout}",
  title="Vehicle Re-Identification for Automatic Video Traffic Surveillance",
  annote="This paper proposes an approach to the vehicle re-identification problem in
a multiple camera system.  We focused on the re-identification itself assuming
that the vehicle detection problem is already solved including extraction of
a full-fledged 3D bounding box. The re-identification problem is solved by using
color histograms and histograms of oriented gradients by a linear regressor.  The
features are used in separate models in order to get the best results in the
shortest CPU computation time. The proposed method works with a high accuracy
(60% true positives retrieved with 10% false positive rate on a challenging
subset of the test data) in 85 milliseconds of the CPU (Core i7) computation time
per one vehicle re-identification assuming the fullHD resolution video input. The
applications of this work include finding important parameters such as travel
time, traffic flow, or traffic information in a distributed traffic surveillance
and monitoring system.",
  address="IEEE Computer Society",
  booktitle="International Workshop on Automatic Traffic Surveillance (CVPR 2016)",
  chapter="130978",
  doi="10.1109/CVPRW.2016.195",
  edition="NEUVEDEN",
  howpublished="online",
  institution="IEEE Computer Society",
  year="2016",
  month="june",
  pages="1568--1574",
  publisher="IEEE Computer Society",
  type="conference paper"
}