Publication detail

Learning Feature Aggregation in Temporal Domain for Re-Identification

ŠPAŇHEL, J. SOCHOR, J. JURÁNEK, R. DOBEŠ, P. BARTL, V. HEROUT, A.

Original Title

Learning Feature Aggregation in Temporal Domain for Re-Identification

English Title

Learning Feature Aggregation in Temporal Domain for Re-Identification

Type

journal article in Web of Science

Language

en

Original Abstract

Person re-identification is a standard and established problem in the computer vision community. In recent years, vehicle re-identification is also getting more attention. In this paper, we focus on both these tasks and propose a method for aggregation of features in temporal domain as it is common to have multiple observations of the same object. The aggregation is based on weighting different elements of the feature vectors by different weights and it is trained in an end-to-end manner by a Siamese network. The experimental results show that our method outperforms other existing methods for feature aggregation in temporal domain on both vehicle and person re-identification tasks. Furthermore, to push research in vehicle re-identification further, we introduce a novel dataset CarsReId74k. The dataset is not limited to frontal/rear viewpoints. It contains 17,681 unique vehicles, 73,976 observed tracks, and 277,236 positive pairs. The dataset was captured by 66 cameras from various angles.

English abstract

Person re-identification is a standard and established problem in the computer vision community. In recent years, vehicle re-identification is also getting more attention. In this paper, we focus on both these tasks and propose a method for aggregation of features in temporal domain as it is common to have multiple observations of the same object. The aggregation is based on weighting different elements of the feature vectors by different weights and it is trained in an end-to-end manner by a Siamese network. The experimental results show that our method outperforms other existing methods for feature aggregation in temporal domain on both vehicle and person re-identification tasks. Furthermore, to push research in vehicle re-identification further, we introduce a novel dataset CarsReId74k. The dataset is not limited to frontal/rear viewpoints. It contains 17,681 unique vehicles, 73,976 observed tracks, and 277,236 positive pairs. The dataset was captured by 66 cameras from various angles.

Keywords

person re-identification, vehicle re-identification, feature aggregation, temporal domain, neural network, traffic surveillance

Released

28.11.2019

Publisher

Elsevier Science

Location

Amsterdam

ISBN

1077-3142

Periodical

COMPUTER VISION AND IMAGE UNDERSTANDING

Year of study

192

Number

11

State

US

Pages from

1

Pages to

12

Pages count

12

URL

Documents

BibTex


@article{BUT161466,
  author="Jakub {Špaňhel} and Jakub {Sochor} and Roman {Juránek} and Petr {Dobeš} and Vojtěch {Bartl} and Adam {Herout}",
  title="Learning Feature Aggregation in Temporal Domain for Re-Identification",
  annote="Person re-identification is a standard and established problem in the computer
vision community. In recent years, vehicle re-identification is also getting more
attention. In this paper, we focus on both these tasks and propose a method for
aggregation of features in temporal domain as it is common to have multiple
observations of the same object. The aggregation is based on weighting different
elements of the feature vectors by different weights and it is trained in an
end-to-end manner by a Siamese network. The experimental results show that our
method outperforms other existing methods for feature aggregation in temporal
domain on both vehicle and person re-identification tasks. Furthermore, to push
research in vehicle re-identification further, we introduce a novel dataset
CarsReId74k. The dataset is not limited to frontal/rear viewpoints. It contains
17,681 unique vehicles, 73,976 observed tracks, and 277,236 positive pairs. The
dataset was captured by 66 cameras from various angles.",
  address="Elsevier Science",
  booktitle="Computer Vision and Image Understanding",
  chapter="161466",
  doi="10.1016/j.cviu.2019.102883",
  edition="NEUVEDEN",
  howpublished="online",
  institution="Elsevier Science",
  number="11",
  volume="192",
  year="2019",
  month="november",
  pages="1--12",
  publisher="Elsevier Science",
  type="journal article in Web of Science"
}