Pedestrian near-miss analysis on vehicle-mounted driving recorders

Teppei Suzuki, Yoshimitsu Aoki, Hirokatsu Kataoka

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Recently, a demand for video analysis on vehicle-mounted driving recorders has been increasing in vision-based safety systems, such as for autonomous vehicles. The technology must be positioned one of the most important task, however, the conventional traffic datasets (e.g. KITTI, Caltech Pedestrian) are not included any dangerous scenes (near-miss scenes), even though the objective of a safety system is to avoid danger. In this paper, (i) we create a pedestrian near-miss dataset on vehicle-mounted driving recorders and (ii) propose a method to jointly learns to predict pedestrian detection and its danger level {high, low, no-danger} with convolutional neural networks (CNN) based on the ResNets. According to the result, we demonstrate the effectiveness of our approach that achieved 68% accuracy of joint pedestrian detection and danger label prediction, and 58.6fps processing time on the self-collected pedestrian near-miss dataset.

Original languageEnglish
Title of host publicationProceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages416-419
Number of pages4
ISBN (Electronic)9784901122160
DOIs
Publication statusPublished - 2017 Jul 19
Event15th IAPR International Conference on Machine Vision Applications, MVA 2017 - Nagoya, Japan
Duration: 2017 May 82017 May 12

Other

Other15th IAPR International Conference on Machine Vision Applications, MVA 2017
CountryJapan
CityNagoya
Period17/5/817/5/12

Fingerprint

Security systems
Labels
Neural networks
Processing

ASJC Scopus subject areas

  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Suzuki, T., Aoki, Y., & Kataoka, H. (2017). Pedestrian near-miss analysis on vehicle-mounted driving recorders. In Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017 (pp. 416-419). [7986889] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.23919/MVA.2017.7986889

Pedestrian near-miss analysis on vehicle-mounted driving recorders. / Suzuki, Teppei; Aoki, Yoshimitsu; Kataoka, Hirokatsu.

Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017. Institute of Electrical and Electronics Engineers Inc., 2017. p. 416-419 7986889.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Suzuki, T, Aoki, Y & Kataoka, H 2017, Pedestrian near-miss analysis on vehicle-mounted driving recorders. in Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017., 7986889, Institute of Electrical and Electronics Engineers Inc., pp. 416-419, 15th IAPR International Conference on Machine Vision Applications, MVA 2017, Nagoya, Japan, 17/5/8. https://doi.org/10.23919/MVA.2017.7986889
Suzuki T, Aoki Y, Kataoka H. Pedestrian near-miss analysis on vehicle-mounted driving recorders. In Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017. Institute of Electrical and Electronics Engineers Inc. 2017. p. 416-419. 7986889 https://doi.org/10.23919/MVA.2017.7986889
Suzuki, Teppei ; Aoki, Yoshimitsu ; Kataoka, Hirokatsu. / Pedestrian near-miss analysis on vehicle-mounted driving recorders. Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 416-419
@inproceedings{c2f635b0140a4a07aaeb6976f218b8ee,
title = "Pedestrian near-miss analysis on vehicle-mounted driving recorders",
abstract = "Recently, a demand for video analysis on vehicle-mounted driving recorders has been increasing in vision-based safety systems, such as for autonomous vehicles. The technology must be positioned one of the most important task, however, the conventional traffic datasets (e.g. KITTI, Caltech Pedestrian) are not included any dangerous scenes (near-miss scenes), even though the objective of a safety system is to avoid danger. In this paper, (i) we create a pedestrian near-miss dataset on vehicle-mounted driving recorders and (ii) propose a method to jointly learns to predict pedestrian detection and its danger level {high, low, no-danger} with convolutional neural networks (CNN) based on the ResNets. According to the result, we demonstrate the effectiveness of our approach that achieved 68{\%} accuracy of joint pedestrian detection and danger label prediction, and 58.6fps processing time on the self-collected pedestrian near-miss dataset.",
author = "Teppei Suzuki and Yoshimitsu Aoki and Hirokatsu Kataoka",
year = "2017",
month = "7",
day = "19",
doi = "10.23919/MVA.2017.7986889",
language = "English",
pages = "416--419",
booktitle = "Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Pedestrian near-miss analysis on vehicle-mounted driving recorders

AU - Suzuki, Teppei

AU - Aoki, Yoshimitsu

AU - Kataoka, Hirokatsu

PY - 2017/7/19

Y1 - 2017/7/19

N2 - Recently, a demand for video analysis on vehicle-mounted driving recorders has been increasing in vision-based safety systems, such as for autonomous vehicles. The technology must be positioned one of the most important task, however, the conventional traffic datasets (e.g. KITTI, Caltech Pedestrian) are not included any dangerous scenes (near-miss scenes), even though the objective of a safety system is to avoid danger. In this paper, (i) we create a pedestrian near-miss dataset on vehicle-mounted driving recorders and (ii) propose a method to jointly learns to predict pedestrian detection and its danger level {high, low, no-danger} with convolutional neural networks (CNN) based on the ResNets. According to the result, we demonstrate the effectiveness of our approach that achieved 68% accuracy of joint pedestrian detection and danger label prediction, and 58.6fps processing time on the self-collected pedestrian near-miss dataset.

AB - Recently, a demand for video analysis on vehicle-mounted driving recorders has been increasing in vision-based safety systems, such as for autonomous vehicles. The technology must be positioned one of the most important task, however, the conventional traffic datasets (e.g. KITTI, Caltech Pedestrian) are not included any dangerous scenes (near-miss scenes), even though the objective of a safety system is to avoid danger. In this paper, (i) we create a pedestrian near-miss dataset on vehicle-mounted driving recorders and (ii) propose a method to jointly learns to predict pedestrian detection and its danger level {high, low, no-danger} with convolutional neural networks (CNN) based on the ResNets. According to the result, we demonstrate the effectiveness of our approach that achieved 68% accuracy of joint pedestrian detection and danger label prediction, and 58.6fps processing time on the self-collected pedestrian near-miss dataset.

UR - http://www.scopus.com/inward/record.url?scp=85027854390&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85027854390&partnerID=8YFLogxK

U2 - 10.23919/MVA.2017.7986889

DO - 10.23919/MVA.2017.7986889

M3 - Conference contribution

SP - 416

EP - 419

BT - Proceedings of the 15th IAPR International Conference on Machine Vision Applications, MVA 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -