Multi-Perspective Vehicle Detection and Tracking: Challenges, Dataset, and Metrics

Jacob Velling Dueholm, Miklas Strøm Kristoffersen, Ravi K. Satzoda, Eshed Ohn-Bar, Thomas B. Moeslund, Mohan M. Trivedi

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

6 Citations (Scopus)

Abstract

The research community has shown significant improvements
in both vision-based detection and tracking of vehicles,
working towards a high level understanding of on-road
maneuvers. Behaviors of surrounding vehicles in a highway
environment is found as an interesting starting point, of
why this dataset is introduced along with its challenges
and evaluation metrics. A vision-based multi-perspective
dataset is presented, containing a full panoramic view from
a moving platform driving on U.S. highways capturing
2704x1440 resolution images at 12 frames per second. The
dataset serves multiple purposes to be used as traditional
detection and tracking, together with tracking of vehicles
across perspectives. Each of the four perspectives have
been annotated, resulting in more than 4000 bounding boxes
in order to evaluate and compare novel methods.
Original languageEnglish
Title of host publication2016 IEEE 19th International Conference on Intelligent Transportation Systems
PublisherIEEE
Publication date26 Dec 2016
Pages959-964
Article number575
ISBN (Electronic)978-1-5090-1889-5
DOIs
Publication statusPublished - 26 Dec 2016
EventInternational Conference on Intelligent Transportation Systems - Convention Center, Windsor Oceanico Hotel, Rio de Janeiro, Brazil
Duration: 1 Nov 20164 Nov 2016
Conference number: 19

Conference

ConferenceInternational Conference on Intelligent Transportation Systems
Number19
LocationConvention Center, Windsor Oceanico Hotel
CountryBrazil
CityRio de Janeiro
Period01/11/201604/11/2016

Fingerprint

Image resolution

Cite this

Dueholm, J. V., Kristoffersen, M. S., Satzoda, R. K., Ohn-Bar, E., Moeslund, T. B., & Trivedi, M. M. (2016). Multi-Perspective Vehicle Detection and Tracking: Challenges, Dataset, and Metrics. In 2016 IEEE 19th International Conference on Intelligent Transportation Systems (pp. 959-964). [575] IEEE. https://doi.org/10.1109/ITSC.2016.7795671
Dueholm, Jacob Velling ; Kristoffersen, Miklas Strøm ; Satzoda, Ravi K. ; Ohn-Bar, Eshed ; Moeslund, Thomas B. ; Trivedi, Mohan M. / Multi-Perspective Vehicle Detection and Tracking : Challenges, Dataset, and Metrics. 2016 IEEE 19th International Conference on Intelligent Transportation Systems. IEEE, 2016. pp. 959-964
@inproceedings{ca927974d6f94728a1959d85469431be,
title = "Multi-Perspective Vehicle Detection and Tracking: Challenges, Dataset, and Metrics",
abstract = "The research community has shown significant improvementsin both vision-based detection and tracking of vehicles,working towards a high level understanding of on-roadmaneuvers. Behaviors of surrounding vehicles in a highwayenvironment is found as an interesting starting point, ofwhy this dataset is introduced along with its challengesand evaluation metrics. A vision-based multi-perspectivedataset is presented, containing a full panoramic view froma moving platform driving on U.S. highways capturing2704x1440 resolution images at 12 frames per second. Thedataset serves multiple purposes to be used as traditionaldetection and tracking, together with tracking of vehiclesacross perspectives. Each of the four perspectives havebeen annotated, resulting in more than 4000 bounding boxesin order to evaluate and compare novel methods.",
author = "Dueholm, {Jacob Velling} and Kristoffersen, {Miklas Str{\o}m} and Satzoda, {Ravi K.} and Eshed Ohn-Bar and Moeslund, {Thomas B.} and Trivedi, {Mohan M.}",
year = "2016",
month = "12",
day = "26",
doi = "10.1109/ITSC.2016.7795671",
language = "English",
pages = "959--964",
booktitle = "2016 IEEE 19th International Conference on Intelligent Transportation Systems",
publisher = "IEEE",
address = "United States",

}

Dueholm, JV, Kristoffersen, MS, Satzoda, RK, Ohn-Bar, E, Moeslund, TB & Trivedi, MM 2016, Multi-Perspective Vehicle Detection and Tracking: Challenges, Dataset, and Metrics. in 2016 IEEE 19th International Conference on Intelligent Transportation Systems., 575, IEEE, pp. 959-964, International Conference on Intelligent Transportation Systems, Rio de Janeiro, Brazil, 01/11/2016. https://doi.org/10.1109/ITSC.2016.7795671

Multi-Perspective Vehicle Detection and Tracking : Challenges, Dataset, and Metrics. / Dueholm, Jacob Velling; Kristoffersen, Miklas Strøm; Satzoda, Ravi K.; Ohn-Bar, Eshed; Moeslund, Thomas B.; Trivedi, Mohan M.

2016 IEEE 19th International Conference on Intelligent Transportation Systems. IEEE, 2016. p. 959-964 575.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Multi-Perspective Vehicle Detection and Tracking

T2 - Challenges, Dataset, and Metrics

AU - Dueholm, Jacob Velling

AU - Kristoffersen, Miklas Strøm

AU - Satzoda, Ravi K.

AU - Ohn-Bar, Eshed

AU - Moeslund, Thomas B.

AU - Trivedi, Mohan M.

PY - 2016/12/26

Y1 - 2016/12/26

N2 - The research community has shown significant improvementsin both vision-based detection and tracking of vehicles,working towards a high level understanding of on-roadmaneuvers. Behaviors of surrounding vehicles in a highwayenvironment is found as an interesting starting point, ofwhy this dataset is introduced along with its challengesand evaluation metrics. A vision-based multi-perspectivedataset is presented, containing a full panoramic view froma moving platform driving on U.S. highways capturing2704x1440 resolution images at 12 frames per second. Thedataset serves multiple purposes to be used as traditionaldetection and tracking, together with tracking of vehiclesacross perspectives. Each of the four perspectives havebeen annotated, resulting in more than 4000 bounding boxesin order to evaluate and compare novel methods.

AB - The research community has shown significant improvementsin both vision-based detection and tracking of vehicles,working towards a high level understanding of on-roadmaneuvers. Behaviors of surrounding vehicles in a highwayenvironment is found as an interesting starting point, ofwhy this dataset is introduced along with its challengesand evaluation metrics. A vision-based multi-perspectivedataset is presented, containing a full panoramic view froma moving platform driving on U.S. highways capturing2704x1440 resolution images at 12 frames per second. Thedataset serves multiple purposes to be used as traditionaldetection and tracking, together with tracking of vehiclesacross perspectives. Each of the four perspectives havebeen annotated, resulting in more than 4000 bounding boxesin order to evaluate and compare novel methods.

U2 - 10.1109/ITSC.2016.7795671

DO - 10.1109/ITSC.2016.7795671

M3 - Article in proceeding

SP - 959

EP - 964

BT - 2016 IEEE 19th International Conference on Intelligent Transportation Systems

PB - IEEE

ER -

Dueholm JV, Kristoffersen MS, Satzoda RK, Ohn-Bar E, Moeslund TB, Trivedi MM. Multi-Perspective Vehicle Detection and Tracking: Challenges, Dataset, and Metrics. In 2016 IEEE 19th International Conference on Intelligent Transportation Systems. IEEE. 2016. p. 959-964. 575 https://doi.org/10.1109/ITSC.2016.7795671