Real-time 3D multi-person pose estimation using an omnidirectional camera and mmWave radars
Real-time 3D multi-person pose estimation using an omnidirectional camera and mmWave radars
Learning-based monocular 3D human pose estimation holds significant potential for a variety of applications, including sports, automation, and entertainment; however, not always at a cost that allows it to be scaled. This paper proposes an affordable solution to learning-based monocular 3D pose estimation from 2D videos that can be utilised outdoors and indoors. We introduce a system that leverages an omnidirectional camera and mmWave radars to estimate the 3D pose of the people in the scene in real-time. The proposed algorithm shows good pose reconstruction accuracy with the average Euclidean distance between a ground truth body joint position and its 3D reconstruction ranging from 4.5cm to 19cm within 20 meters along both the x and z axes of the camera.
Human Pose Estimation, Omnidirectional camera, Radar Sensors, Real-Time System
Amin, Aarti
ee8f1a8d-3f93-4037-85e4-713d8c1ba78e
Tamajo, Alberto
190a4680-7aa4-4d37-82b9-e03b96c5965f
Klugman, Isaac
06ba127e-c59c-49f7-83d0-0db857d51d41
Stoev, Emil
8a095f89-4600-4909-a867-9fe431a48884
Fisho, Timothy
287a774a-4031-476f-ab5f-5573ddef2ed1
Lim, Hwasup
3c9f6a38-639d-4c4c-afc2-a0704166d72e
Kim, Hansung
2c7c135c-f00b-4409-acb2-85b3a9e8225f
2023
Amin, Aarti
ee8f1a8d-3f93-4037-85e4-713d8c1ba78e
Tamajo, Alberto
190a4680-7aa4-4d37-82b9-e03b96c5965f
Klugman, Isaac
06ba127e-c59c-49f7-83d0-0db857d51d41
Stoev, Emil
8a095f89-4600-4909-a867-9fe431a48884
Fisho, Timothy
287a774a-4031-476f-ab5f-5573ddef2ed1
Lim, Hwasup
3c9f6a38-639d-4c4c-afc2-a0704166d72e
Kim, Hansung
2c7c135c-f00b-4409-acb2-85b3a9e8225f
Amin, Aarti, Tamajo, Alberto, Klugman, Isaac, Stoev, Emil, Fisho, Timothy, Lim, Hwasup and Kim, Hansung
(2023)
Real-time 3D multi-person pose estimation using an omnidirectional camera and mmWave radars.
In 2023 International Conference on Engineering and Emerging Technologies (ICEET).
IEEE.
6 pp
.
(doi:10.1109/ICEET60227.2023.10526102).
Record type:
Conference or Workshop Item
(Paper)
Abstract
Learning-based monocular 3D human pose estimation holds significant potential for a variety of applications, including sports, automation, and entertainment; however, not always at a cost that allows it to be scaled. This paper proposes an affordable solution to learning-based monocular 3D pose estimation from 2D videos that can be utilised outdoors and indoors. We introduce a system that leverages an omnidirectional camera and mmWave radars to estimate the 3D pose of the people in the scene in real-time. The proposed algorithm shows good pose reconstruction accuracy with the average Euclidean distance between a ground truth body joint position and its 3D reconstruction ranging from 4.5cm to 19cm within 20 meters along both the x and z axes of the camera.
Text
124_23ICEET-Aarti
- Accepted Manuscript
Restricted to Repository staff only until 17 May 2026.
Request a copy
More information
Published date: 2023
Additional Information:
Publisher Copyright:
© 2023 IEEE.
Venue - Dates:
International Conference on Engineering and Emerging Technologies, Istanbul Topkapi University, İstanbul, Turkey, 2023-10-27 - 2023-10-28
Keywords:
Human Pose Estimation, Omnidirectional camera, Radar Sensors, Real-Time System
Identifiers
Local EPrints ID: 490635
URI: http://eprints.soton.ac.uk/id/eprint/490635
PURE UUID: a806380c-4249-4d6c-a8ea-5be2bbbe1c5e
Catalogue record
Date deposited: 31 May 2024 16:49
Last modified: 16 Jul 2024 02:06
Export record
Altmetrics
Contributors
Author:
Aarti Amin
Author:
Alberto Tamajo
Author:
Isaac Klugman
Author:
Emil Stoev
Author:
Timothy Fisho
Author:
Hwasup Lim
Author:
Hansung Kim
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics