Lightweight deep learning for real-time road distress detection on mobile devices

Hu, Yuanyuan; Chen, Ning; Hou, Yue; Lin, Xingshi; Jing, Baohong; Liu, Pengfei

doi:10.1038/s41467-025-59516-5

Journal article 146 views 34 downloads

Lightweight deep learning for real-time road distress detection on mobile devices

Yuanyuan Hu

, Ning Chen, Yue Hou

, Xingshi Lin, Baohong Jing, Pengfei Liu

Nature Communications, Volume: 16, Start page: 4212

Swansea University Author: Yue Hou

PDF | Version of Record

© The Author(s) 2025. This article is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).
Download (3.23MB)

Check full text

DOI (Published version): 10.1038/s41467-025-59516-5

Abstract

Efficient and accurate road distress detection is crucial for infrastructure maintenance and transportation safety. Traditional manual inspections are labor-intensive and time-consuming, while increasingly popular automated systems often rely on computationally intensive devices, limiting widespread...

Full description

Published in:	Nature Communications
ISSN:	2041-1723
Published:	Springer Nature 2025
Online Access:	Check full text
URI:	https://cronfa.swan.ac.uk/Record/cronfa69449

first_indexed	2025-05-08T09:50:32Z
last_indexed	2025-05-09T07:04:49Z
id	cronfa69449
recordtype	SURis
fullrecord	<?xml version="1.0"?><rfc1807><datestamp>2025-05-08T10:55:14.2072001</datestamp><bib-version>v2</bib-version><id>69449</id><entry>2025-05-08</entry><title>Lightweight deep learning for real-time road distress detection on mobile devices</title><swanseaauthors><author><sid>92bf566c65343cb3ee04ad963eacf31b</sid><ORCID>0000-0002-4334-2620</ORCID><firstname>Yue</firstname><surname>Hou</surname><name>Yue Hou</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2025-05-08</date><deptcode>ACEM</deptcode><abstract>Efficient and accurate road distress detection is crucial for infrastructure maintenance and transportation safety. Traditional manual inspections are labor-intensive and time-consuming, while increasingly popular automated systems often rely on computationally intensive devices, limiting widespread adoption. To address these challenges, this study introduces MobiLiteNet, a lightweight deep learning approach designed for mobile deployment on smartphones and mixed reality systems. Utilizing a diverse dataset collected from Europe and Asia, MobiLiteNet incorporates Efficient Channel Attention to boost model performance, followed by structural refinement, sparse knowledge distillation, structured pruning, and quantization to significantly increase the computational efficiency while preserving high detection accuracy. To validate its effectiveness, MobiLiteNet improves the existing MobileNet model. Test results show that the improved MobileNet outperforms baseline models on mobile devices. With significantly reduced computational costs, this approach enables real-time, scalable, and accurate road distress detection, contributing to more efficient road infrastructure management and intelligent transportation systems.</abstract><type>Journal Article</type><journal>Nature Communications</journal><volume>16</volume><journalNumber/><paginationStart>4212</paginationStart><paginationEnd/><publisher>Springer Nature</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic>2041-1723</issnElectronic><keywords/><publishedDay>6</publishedDay><publishedMonth>5</publishedMonth><publishedYear>2025</publishedYear><publishedDate>2025-05-06</publishedDate><doi>10.1038/s41467-025-59516-5</doi><url/><notes/><college>COLLEGE NANME</college><department>Aerospace, Civil, Electrical, and Mechanical Engineering</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>ACEM</DepartmentCode><institution>Swansea University</institution><apcterm>Another institution paid the OA fee</apcterm><funders>Open Access funding enabled and organized by Projekt DEAL.</funders><projectreference/><lastEdited>2025-05-08T10:55:14.2072001</lastEdited><Created>2025-05-08T09:40:49.8571437</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering</level></path><authors><author><firstname>Yuanyuan</firstname><surname>Hu</surname><orcid>0009-0002-3653-4947</orcid><order>1</order></author><author><firstname>Ning</firstname><surname>Chen</surname><order>2</order></author><author><firstname>Yue</firstname><surname>Hou</surname><orcid>0000-0002-4334-2620</orcid><order>3</order></author><author><firstname>Xingshi</firstname><surname>Lin</surname><order>4</order></author><author><firstname>Baohong</firstname><surname>Jing</surname><order>5</order></author><author><firstname>Pengfei</firstname><surname>Liu</surname><orcid>0000-0001-5983-7305</orcid><order>6</order></author></authors><documents><document><filename>69449__34200__2496fc1aa5be41eca551927780d33d63.pdf</filename><originalFilename>69449.VOR.pdf</originalFilename><uploaded>2025-05-08T10:47:56.9334469</uploaded><type>Output</type><contentLength>3388554</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>© The Author(s) 2025. This article is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>http://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling	2025-05-08T10:55:14.2072001 v2 69449 2025-05-08 Lightweight deep learning for real-time road distress detection on mobile devices 92bf566c65343cb3ee04ad963eacf31b 0000-0002-4334-2620 Yue Hou Yue Hou true false 2025-05-08 ACEM Efficient and accurate road distress detection is crucial for infrastructure maintenance and transportation safety. Traditional manual inspections are labor-intensive and time-consuming, while increasingly popular automated systems often rely on computationally intensive devices, limiting widespread adoption. To address these challenges, this study introduces MobiLiteNet, a lightweight deep learning approach designed for mobile deployment on smartphones and mixed reality systems. Utilizing a diverse dataset collected from Europe and Asia, MobiLiteNet incorporates Efficient Channel Attention to boost model performance, followed by structural refinement, sparse knowledge distillation, structured pruning, and quantization to significantly increase the computational efficiency while preserving high detection accuracy. To validate its effectiveness, MobiLiteNet improves the existing MobileNet model. Test results show that the improved MobileNet outperforms baseline models on mobile devices. With significantly reduced computational costs, this approach enables real-time, scalable, and accurate road distress detection, contributing to more efficient road infrastructure management and intelligent transportation systems. Journal Article Nature Communications 16 4212 Springer Nature 2041-1723 6 5 2025 2025-05-06 10.1038/s41467-025-59516-5 COLLEGE NANME Aerospace, Civil, Electrical, and Mechanical Engineering COLLEGE CODE ACEM Swansea University Another institution paid the OA fee Open Access funding enabled and organized by Projekt DEAL. 2025-05-08T10:55:14.2072001 2025-05-08T09:40:49.8571437 Faculty of Science and Engineering School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering Yuanyuan Hu 0009-0002-3653-4947 1 Ning Chen 2 Yue Hou 0000-0002-4334-2620 3 Xingshi Lin 4 Baohong Jing 5 Pengfei Liu 0000-0001-5983-7305 6 69449__34200__2496fc1aa5be41eca551927780d33d63.pdf 69449.VOR.pdf 2025-05-08T10:47:56.9334469 Output 3388554 application/pdf Version of Record true © The Author(s) 2025. This article is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0). true eng http://creativecommons.org/licenses/by/4.0/
title	Lightweight deep learning for real-time road distress detection on mobile devices
spellingShingle	Lightweight deep learning for real-time road distress detection on mobile devices Yue Hou
title_short	Lightweight deep learning for real-time road distress detection on mobile devices
title_full	Lightweight deep learning for real-time road distress detection on mobile devices
title_fullStr	Lightweight deep learning for real-time road distress detection on mobile devices
title_full_unstemmed	Lightweight deep learning for real-time road distress detection on mobile devices
title_sort	Lightweight deep learning for real-time road distress detection on mobile devices
author_id_str_mv	92bf566c65343cb3ee04ad963eacf31b
author_id_fullname_str_mv	92bf566c65343cb3ee04ad963eacf31b_***_Yue Hou
author	Yue Hou
author2	Yuanyuan Hu Ning Chen Yue Hou Xingshi Lin Baohong Jing Pengfei Liu
format	Journal article
container_title	Nature Communications
container_volume	16
container_start_page	4212
publishDate	2025
institution	Swansea University
issn	2041-1723
doi_str_mv	10.1038/s41467-025-59516-5
publisher	Springer Nature
college_str	Faculty of Science and Engineering
hierarchytype
hierarchy_top_id	facultyofscienceandengineering
hierarchy_top_title	Faculty of Science and Engineering
hierarchy_parent_id	facultyofscienceandengineering
hierarchy_parent_title	Faculty of Science and Engineering
department_str	School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Aerospace, Civil, Electrical, General and Mechanical Engineering - Civil Engineering
document_store_str	1
active_str	0
description	Efficient and accurate road distress detection is crucial for infrastructure maintenance and transportation safety. Traditional manual inspections are labor-intensive and time-consuming, while increasingly popular automated systems often rely on computationally intensive devices, limiting widespread adoption. To address these challenges, this study introduces MobiLiteNet, a lightweight deep learning approach designed for mobile deployment on smartphones and mixed reality systems. Utilizing a diverse dataset collected from Europe and Asia, MobiLiteNet incorporates Efficient Channel Attention to boost model performance, followed by structural refinement, sparse knowledge distillation, structured pruning, and quantization to significantly increase the computational efficiency while preserving high detection accuracy. To validate its effectiveness, MobiLiteNet improves the existing MobileNet model. Test results show that the improved MobileNet outperforms baseline models on mobile devices. With significantly reduced computational costs, this approach enables real-time, scalable, and accurate road distress detection, contributing to more efficient road infrastructure management and intelligent transportation systems.
published_date	2025-05-06T05:55:17Z
_version_	1835784908663947264
score	11.064947

Lightweight deep learning for real-time road distress detection on mobile devices

Similar Items