Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices

Real-time object pose estimation and tracking is challenging but essential for some emerging applications, such as augmented reality. In general, state-of-the-art methods address this problem using deep neural networks, which indeed yield satisfactory results. Nevertheless, the high computational co...

Full description

Saved in:
Bibliographic Details
Main Authors: Yo-Chung Lau, Kuan-Wei Tseng, Peng-Yuan Kao, I-Ju Hsieh, Hsiao-Ching Tseng, Yi-Ping Hung
Format: Article
Language:English
Published: IEEE 2023-01-01
Series:IEEE Journal of Indoor and Seamless Positioning and Navigation
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10352604/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850051513517342720
author Yo-Chung Lau
Kuan-Wei Tseng
Peng-Yuan Kao
I-Ju Hsieh
Hsiao-Ching Tseng
Yi-Ping Hung
author_facet Yo-Chung Lau
Kuan-Wei Tseng
Peng-Yuan Kao
I-Ju Hsieh
Hsiao-Ching Tseng
Yi-Ping Hung
author_sort Yo-Chung Lau
collection DOAJ
description Real-time object pose estimation and tracking is challenging but essential for some emerging applications, such as augmented reality. In general, state-of-the-art methods address this problem using deep neural networks, which indeed yield satisfactory results. Nevertheless, the high computational cost of these methods makes them unsuitable for mobile devices where real-world applications usually take place. We propose real-time object pose tracking system with low computational cost for mobile devices. It is a monocular inertial-assisted-visual system with a client–server architecture connected by high-speed networking. Inertial measurement unit (IMU) pose propagation is performed on the client side for fast pose tracking, and RGB image-based 3-D object pose estimation is performed on the server side to obtain accurate poses, after which the pose is sent to the client side for refinement, where we propose a bias self-correction mechanism to reduce the drift. We also propose a fast and effective pose inspection algorithm to detect tracking failures and incorrect pose estimation. In this way, the pose updates rapidly even within 5 ms on low-level devices, making it possible to support real-time tracking for applications. In addition, an object pose dataset with RGB images and IMU measurements is delivered for evaluation. Experiments also show that our method performs well with both accuracy and robustness.
format Article
id doaj-art-66d31d4dcbd94cd3889efdf0fb4ad04b
institution DOAJ
issn 2832-7322
language English
publishDate 2023-01-01
publisher IEEE
record_format Article
series IEEE Journal of Indoor and Seamless Positioning and Navigation
spelling doaj-art-66d31d4dcbd94cd3889efdf0fb4ad04b2025-08-20T02:53:07ZengIEEEIEEE Journal of Indoor and Seamless Positioning and Navigation2832-73222023-01-01121122010.1109/JISPIN.2023.334098710352604Real-Time Object Pose Tracking System With Low Computational Cost for Mobile DevicesYo-Chung Lau0https://orcid.org/0009-0005-8251-1270Kuan-Wei Tseng1https://orcid.org/0000-0003-1134-5314Peng-Yuan Kao2https://orcid.org/0000-0002-5582-1039I-Ju Hsieh3https://orcid.org/0009-0002-0561-5154Hsiao-Ching Tseng4https://orcid.org/0009-0006-1590-8072Yi-Ping Hung5https://orcid.org/0000-0002-9373-2184Graduate Institute of Networking and Multimedia, National Taiwan University, Taipei, TaiwanDepartment of Computer Science, Tokyo Institute of Technology, Tokyo, JapanGraduate Institute of Networking and Multimedia, National Taiwan University, Taipei, TaiwanDepartment of Computer Science and Information Engineering, National Taiwan University, Taipei, TaiwanDepartment of Computer Science and Information Engineering, National Taiwan University, Taipei, TaiwanGraduate Institute of Networking and Multimedia, National Taiwan University, Taipei, TaiwanReal-time object pose estimation and tracking is challenging but essential for some emerging applications, such as augmented reality. In general, state-of-the-art methods address this problem using deep neural networks, which indeed yield satisfactory results. Nevertheless, the high computational cost of these methods makes them unsuitable for mobile devices where real-world applications usually take place. We propose real-time object pose tracking system with low computational cost for mobile devices. It is a monocular inertial-assisted-visual system with a client–server architecture connected by high-speed networking. Inertial measurement unit (IMU) pose propagation is performed on the client side for fast pose tracking, and RGB image-based 3-D object pose estimation is performed on the server side to obtain accurate poses, after which the pose is sent to the client side for refinement, where we propose a bias self-correction mechanism to reduce the drift. We also propose a fast and effective pose inspection algorithm to detect tracking failures and incorrect pose estimation. In this way, the pose updates rapidly even within 5 ms on low-level devices, making it possible to support real-time tracking for applications. In addition, an object pose dataset with RGB images and IMU measurements is delivered for evaluation. Experiments also show that our method performs well with both accuracy and robustness.https://ieeexplore.ieee.org/document/10352604/Low computational costmobile deviceobject pose estimationreal-time object pose tracking
spellingShingle Yo-Chung Lau
Kuan-Wei Tseng
Peng-Yuan Kao
I-Ju Hsieh
Hsiao-Ching Tseng
Yi-Ping Hung
Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices
IEEE Journal of Indoor and Seamless Positioning and Navigation
Low computational cost
mobile device
object pose estimation
real-time object pose tracking
title Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices
title_full Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices
title_fullStr Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices
title_full_unstemmed Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices
title_short Real-Time Object Pose Tracking System With Low Computational Cost for Mobile Devices
title_sort real time object pose tracking system with low computational cost for mobile devices
topic Low computational cost
mobile device
object pose estimation
real-time object pose tracking
url https://ieeexplore.ieee.org/document/10352604/
work_keys_str_mv AT yochunglau realtimeobjectposetrackingsystemwithlowcomputationalcostformobiledevices
AT kuanweitseng realtimeobjectposetrackingsystemwithlowcomputationalcostformobiledevices
AT pengyuankao realtimeobjectposetrackingsystemwithlowcomputationalcostformobiledevices
AT ijuhsieh realtimeobjectposetrackingsystemwithlowcomputationalcostformobiledevices
AT hsiaochingtseng realtimeobjectposetrackingsystemwithlowcomputationalcostformobiledevices
AT yipinghung realtimeobjectposetrackingsystemwithlowcomputationalcostformobiledevices