Radar Multi Object Tracking using DNN Features

Mujtaba Hassan; Francesco Fioranelli; Alexander Yarovoy; Satish Ravindran

doi:10.1109/RADAR54928.2023.10371032

Radar Multi Object Tracking using DNN Features

Mujtaba Hassan^*, Francesco Fioranelli, Alexander Yarovoy, Satish Ravindran

^*Corresponding author for this work

Microwave Sensing, Signals & Systems

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

25 Downloads (Pure)

Abstract

A single frame radar-based multi-object tracker that aims to improve data association for better tracking performance is proposed. Firstly, a baseline tracker based on track-by-detection paradigm was implemented for automotive radar. Secondly, investigation on the performance of the tracker when tracking individual classes separately versus all classes together was performed. Thirdly, appearance features were extracted from a neural network and added as an additional metric to the cost matrix for improved data association. Extensive experiments on the 2D RadarScenes dataset and a 3D proprietary Lunewave dataset (in partnership with NXP Semiconductors) showed a consistent improvement in the tracking performance using the approach proposed by adding features extracted from a neural network.

Original language	English
Title of host publication	2023 IEEE International Radar Conference, RADAR 2023
Publisher	IEEE
Number of pages	6
ISBN (Electronic)	9781665482783
DOIs	https://doi.org/10.1109/RADAR54928.2023.10371032
Publication status	Published - 2023
Event	2023 IEEE International Radar Conference, RADAR 2023 - Sydney, Australia Duration: 6 Nov 2023 → 10 Nov 2023

Publication series

Name	Proceedings of the IEEE Radar Conference
ISSN (Print)	1097-5764
ISSN (Electronic)	2375-5318

Conference

Conference	2023 IEEE International Radar Conference, RADAR 2023
Country/Territory	Australia
City	Sydney
Period	6/11/23 → 10/11/23

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

Data association
detector
track-by-detection

Access to Document

10.1109/RADAR54928.2023.10371032

Radar_Multi_Object_Tracking_using_DNN_FeaturesFinal published version, 762 KB

Cite this

@inproceedings{86c066a382824481b6e91665695e6783,

title = "Radar Multi Object Tracking using DNN Features",

abstract = "A single frame radar-based multi-object tracker that aims to improve data association for better tracking performance is proposed. Firstly, a baseline tracker based on track-by-detection paradigm was implemented for automotive radar. Secondly, investigation on the performance of the tracker when tracking individual classes separately versus all classes together was performed. Thirdly, appearance features were extracted from a neural network and added as an additional metric to the cost matrix for improved data association. Extensive experiments on the 2D RadarScenes dataset and a 3D proprietary Lunewave dataset (in partnership with NXP Semiconductors) showed a consistent improvement in the tracking performance using the approach proposed by adding features extracted from a neural network.",

keywords = "Data association, detector, track-by-detection",

author = "Mujtaba Hassan and Francesco Fioranelli and Alexander Yarovoy and Satish Ravindran",

note = "Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. ; 2023 IEEE International Radar Conference, RADAR 2023 ; Conference date: 06-11-2023 Through 10-11-2023",

year = "2023",

doi = "10.1109/RADAR54928.2023.10371032",

language = "English",

series = "Proceedings of the IEEE Radar Conference",

publisher = "IEEE",

booktitle = "2023 IEEE International Radar Conference, RADAR 2023",

address = "United States",

}

TY - GEN

T1 - Radar Multi Object Tracking using DNN Features

AU - Hassan, Mujtaba

AU - Fioranelli, Francesco

AU - Yarovoy, Alexander

AU - Ravindran, Satish

N1 - Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2023

Y1 - 2023

N2 - A single frame radar-based multi-object tracker that aims to improve data association for better tracking performance is proposed. Firstly, a baseline tracker based on track-by-detection paradigm was implemented for automotive radar. Secondly, investigation on the performance of the tracker when tracking individual classes separately versus all classes together was performed. Thirdly, appearance features were extracted from a neural network and added as an additional metric to the cost matrix for improved data association. Extensive experiments on the 2D RadarScenes dataset and a 3D proprietary Lunewave dataset (in partnership with NXP Semiconductors) showed a consistent improvement in the tracking performance using the approach proposed by adding features extracted from a neural network.

AB - A single frame radar-based multi-object tracker that aims to improve data association for better tracking performance is proposed. Firstly, a baseline tracker based on track-by-detection paradigm was implemented for automotive radar. Secondly, investigation on the performance of the tracker when tracking individual classes separately versus all classes together was performed. Thirdly, appearance features were extracted from a neural network and added as an additional metric to the cost matrix for improved data association. Extensive experiments on the 2D RadarScenes dataset and a 3D proprietary Lunewave dataset (in partnership with NXP Semiconductors) showed a consistent improvement in the tracking performance using the approach proposed by adding features extracted from a neural network.

KW - Data association

KW - detector

KW - track-by-detection

UR - http://www.scopus.com/inward/record.url?scp=85182744085&partnerID=8YFLogxK

U2 - 10.1109/RADAR54928.2023.10371032

DO - 10.1109/RADAR54928.2023.10371032

M3 - Conference contribution

AN - SCOPUS:85182744085

T3 - Proceedings of the IEEE Radar Conference

BT - 2023 IEEE International Radar Conference, RADAR 2023

PB - IEEE

T2 - 2023 IEEE International Radar Conference, RADAR 2023

Y2 - 6 November 2023 through 10 November 2023

ER -

Radar Multi Object Tracking using DNN Features

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this