Real-time object detection with automatic switching between single-board computers and the cloud

Ahmed Osman; Raouf Abozariba; A. Taufiq Asyhari; Adel Aneiba; Ambreen Hussain; Bidushi Barua; Moazam Azeem

doi:10.1109/SSCI50451.2021.9660166

Real-time object detection with automatic switching between single-board computers and the cloud

Ahmed Osman, Raouf Abozariba, A. Taufiq Asyhari, Adel Aneiba, Ambreen Hussain, Bidushi Barua, Moazam Azeem

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research

4 Citations (Scopus)

Abstract

We present a wireless real-time object detection system utilizing single-board devices, cloud computing platforms and web-streaming. Currently, most inference applications statically perform tasks either on local machines or remote cloud servers. However, devices connected through cellular technologies face volatile network conditions, compromising detection performance. Furthermore, while the limited computing power of single-board computers degrade detection correctness, excessive power consumption of machine learning models used for inference reduces operation time. In this paper, we propose a dynamic system that monitors embedded device's wireless link quality and battery level to decide on detecting objects locally or remotely. The experimental results show that our dynamic offloading approach could reduce devices' energy usage while achieving high accuracy, real-time object detection.

Original language	English
Title of host publication	2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings
Editors	Dipti Srinivasan, Anna Wilbik
Place of Publication	Piscataway NJ USA
Publisher	IEEE, Institute of Electrical and Electronics Engineers
Number of pages	6
ISBN (Electronic)	9781728190488
ISBN (Print)	9781728190495
DOIs	https://doi.org/10.1109/SSCI50451.2021.9660166
Publication status	Published - 2021
Externally published	Yes
Event	IEEE Symposium Series on Computational Intelligence (SSCI) 2021 - Orlando, United States of America Duration: 5 Dec 2021 → 7 Dec 2021 https://ieeexplore.ieee.org/xpl/conhome/9659537/proceeding (Proceedings) https://ieeessci2021.org/ (Website)

Conference

Conference	IEEE Symposium Series on Computational Intelligence (SSCI) 2021
Abbreviated title	SSCI 2021
Country/Territory	United States of America
City	Orlando
Period	5/12/21 → 7/12/21
Internet address	https://ieeexplore.ieee.org/xpl/conhome/9659537/proceeding (Proceedings) https://ieeessci2021.org/ (Website)

Keywords

Machine learning
Object detection
WebRTC

Access to Document

10.1109/SSCI50451.2021.9660166

Cite this

Osman, A., Abozariba, R., Asyhari, A. T., Aneiba, A., Hussain, A., Barua, B., & Azeem, M. (2021). Real-time object detection with automatic switching between single-board computers and the cloud. In D. Srinivasan, & A. Wilbik (Eds.), 2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings IEEE, Institute of Electrical and Electronics Engineers. https://doi.org/10.1109/SSCI50451.2021.9660166

@inproceedings{18eea00749ba44e1954dc81342244af0,

title = "Real-time object detection with automatic switching between single-board computers and the cloud",

abstract = "We present a wireless real-time object detection system utilizing single-board devices, cloud computing platforms and web-streaming. Currently, most inference applications statically perform tasks either on local machines or remote cloud servers. However, devices connected through cellular technologies face volatile network conditions, compromising detection performance. Furthermore, while the limited computing power of single-board computers degrade detection correctness, excessive power consumption of machine learning models used for inference reduces operation time. In this paper, we propose a dynamic system that monitors embedded device's wireless link quality and battery level to decide on detecting objects locally or remotely. The experimental results show that our dynamic offloading approach could reduce devices' energy usage while achieving high accuracy, real-time object detection.",

keywords = "Machine learning, Object detection, WebRTC",

author = "Ahmed Osman and Raouf Abozariba and Asyhari, {A. Taufiq} and Adel Aneiba and Ambreen Hussain and Bidushi Barua and Moazam Azeem",

note = "Funding Information: This work was funded by the Department for Digital, Culture, Media & Sport (DCMS) as part of the 5G Connected Forest (5GCF) Project under its 5G Testbeds and Trials Program. Publisher Copyright: {\textcopyright} 2021 IEEE.; IEEE Symposium Series on Computational Intelligence (SSCI) 2021, SSCI 2021 ; Conference date: 05-12-2021 Through 07-12-2021",

year = "2021",

doi = "10.1109/SSCI50451.2021.9660166",

language = "English",

isbn = "9781728190495",

editor = "Dipti Srinivasan and Anna Wilbik",

booktitle = "2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

address = "United States of America",

url = "https://ieeexplore.ieee.org/xpl/conhome/9659537/proceeding, https://ieeessci2021.org/",

}

Osman, A, Abozariba, R, Asyhari, AT, Aneiba, A, Hussain, A, Barua, B & Azeem, M 2021, Real-time object detection with automatic switching between single-board computers and the cloud. in D Srinivasan & A Wilbik (eds), 2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings. IEEE, Institute of Electrical and Electronics Engineers, Piscataway NJ USA, IEEE Symposium Series on Computational Intelligence (SSCI) 2021, Orlando, Florida, United States of America, 5/12/21. https://doi.org/10.1109/SSCI50451.2021.9660166

Real-time object detection with automatic switching between single-board computers and the cloud. / Osman, Ahmed; Abozariba, Raouf; Asyhari, A. Taufiq et al.
2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings. ed. / Dipti Srinivasan; Anna Wilbik. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers, 2021.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research

TY - GEN

T1 - Real-time object detection with automatic switching between single-board computers and the cloud

AU - Osman, Ahmed

AU - Abozariba, Raouf

AU - Asyhari, A. Taufiq

AU - Aneiba, Adel

AU - Hussain, Ambreen

AU - Barua, Bidushi

AU - Azeem, Moazam

N1 - Funding Information: This work was funded by the Department for Digital, Culture, Media & Sport (DCMS) as part of the 5G Connected Forest (5GCF) Project under its 5G Testbeds and Trials Program. Publisher Copyright: © 2021 IEEE.

PY - 2021

Y1 - 2021

N2 - We present a wireless real-time object detection system utilizing single-board devices, cloud computing platforms and web-streaming. Currently, most inference applications statically perform tasks either on local machines or remote cloud servers. However, devices connected through cellular technologies face volatile network conditions, compromising detection performance. Furthermore, while the limited computing power of single-board computers degrade detection correctness, excessive power consumption of machine learning models used for inference reduces operation time. In this paper, we propose a dynamic system that monitors embedded device's wireless link quality and battery level to decide on detecting objects locally or remotely. The experimental results show that our dynamic offloading approach could reduce devices' energy usage while achieving high accuracy, real-time object detection.

AB - We present a wireless real-time object detection system utilizing single-board devices, cloud computing platforms and web-streaming. Currently, most inference applications statically perform tasks either on local machines or remote cloud servers. However, devices connected through cellular technologies face volatile network conditions, compromising detection performance. Furthermore, while the limited computing power of single-board computers degrade detection correctness, excessive power consumption of machine learning models used for inference reduces operation time. In this paper, we propose a dynamic system that monitors embedded device's wireless link quality and battery level to decide on detecting objects locally or remotely. The experimental results show that our dynamic offloading approach could reduce devices' energy usage while achieving high accuracy, real-time object detection.

KW - Machine learning

KW - Object detection

KW - WebRTC

UR - http://www.scopus.com/inward/record.url?scp=85125806519&partnerID=8YFLogxK

U2 - 10.1109/SSCI50451.2021.9660166

DO - 10.1109/SSCI50451.2021.9660166

M3 - Conference Paper

AN - SCOPUS:85125806519

SN - 9781728190495

BT - 2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings

A2 - Srinivasan, Dipti

A2 - Wilbik, Anna

PB - IEEE, Institute of Electrical and Electronics Engineers

CY - Piscataway NJ USA

T2 - IEEE Symposium Series on Computational Intelligence (SSCI) 2021

Y2 - 5 December 2021 through 7 December 2021

ER -

Osman A, Abozariba R, Asyhari AT, Aneiba A, Hussain A, Barua B et al. Real-time object detection with automatic switching between single-board computers and the cloud. In Srinivasan D, Wilbik A, editors, 2021 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings. Piscataway NJ USA: IEEE, Institute of Electrical and Electronics Engineers. 2021 doi: 10.1109/SSCI50451.2021.9660166

Real-time object detection with automatic switching between single-board computers and the cloud

Abstract

Conference

Keywords

Access to Document

Other files and links

Cite this