FMSD: Focal Multi-scale Shape-feature Distillation Network for Small Fasteners Detection in Electric Power Scene
In the electric power scene, fasteners play a pivotal role in securing and connecting electrical equipment, with small fastener detection (SFD) being crucial for ensuring operational stability. Despite the replacement of manual inspection methods by non-destructive techniques employing deep learning...
Gespeichert in:
Veröffentlicht in: | IEEE transactions on circuits and systems for video technology 2024-10, p.1-1 |
---|---|
Hauptverfasser: | , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | 1 |
---|---|
container_issue | |
container_start_page | 1 |
container_title | IEEE transactions on circuits and systems for video technology |
container_volume | |
creator | Yi, Junfei Mao, Jianxu Zhang, Hui Li, Mingjie Zeng, Kai Feng, Mingtao Chang, Xiaojun Wang, Yaonan |
description | In the electric power scene, fasteners play a pivotal role in securing and connecting electrical equipment, with small fastener detection (SFD) being crucial for ensuring operational stability. Despite the replacement of manual inspection methods by non-destructive techniques employing deep learning, these approaches often demand substantial computational resources and involve numerous parameters. While knowledge distillation (KD) can be a viable solution, existing KD methods may often fail to achieve satisfactory performance when dealing with small object presentation and little inter-class variability in SFD tasks. To alleviate this, we propose a Focal Multi-scale Shape-feature Distillation Network (FMSD) to achieve efficient and precise fastener detection in electric power scenarios. Specifically, we propose a novel Multi-Scale Shape-Aware Feature Aggregation module (MSFA) to augment the network's perception of object shape and scale during the KD process. Additionally, we propose a Contour-Guided Distillation (CGD) module to optimize the transfer of the extracted shape-sensitive knowledge between the teacher and student models. Through a series of experiments compared with existing state-of-the-art (SOTA) methods, our method demonstrates superior performance over existing SOTA techniques, both efficiently and effectively. Furthermore, validation on publicly available power scene datasets confirms the generalizability and adaptability of our proposed FMSD across various settings. |
doi_str_mv | 10.1109/TCSVT.2024.3485548 |
format | Article |
fullrecord | <record><control><sourceid>crossref_RIE</sourceid><recordid>TN_cdi_crossref_primary_10_1109_TCSVT_2024_3485548</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10731891</ieee_id><sourcerecordid>10_1109_TCSVT_2024_3485548</sourcerecordid><originalsourceid>FETCH-LOGICAL-c641-7d2d971fdf666da93eac6c340f3ec57e8f9d2bcf70f777eb1ca4475580f4f0ff3</originalsourceid><addsrcrecordid>eNpNkMFKAzEURYMoWKs_IC7yA6lJJpnMuJO2o0KrwgxuhzTzgtG0U5KU4t87tV24endxz4V3ELpldMIYLe-baf3RTDjlYpKJQkpRnKERk7IgnFN5PmQqGSk4k5foKsYvSpkohBqhbbWsZw-46o32eLnzyZE4RMD1p94CsaDTLgCeuZic9zq5foNfIe378I1tH3C91t7jSscEGwgRzyCB-Wu5DZ77IQdn8Hu_h6Frhs41urDaR7g53TFqqnkzfSaLt6eX6eOCmFwwojrelYrZzuZ53ukyA21ykwlqMzBSQWHLjq-MVdQqpWDFjBZCDf9SKyy1Nhsjfpw1oY8xgG23wa11-GkZbQ_K2j9l7UFZe1I2QHdHyAHAP0BlrChZ9gsQJ2on</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>FMSD: Focal Multi-scale Shape-feature Distillation Network for Small Fasteners Detection in Electric Power Scene</title><source>IEEE</source><creator>Yi, Junfei ; Mao, Jianxu ; Zhang, Hui ; Li, Mingjie ; Zeng, Kai ; Feng, Mingtao ; Chang, Xiaojun ; Wang, Yaonan</creator><creatorcontrib>Yi, Junfei ; Mao, Jianxu ; Zhang, Hui ; Li, Mingjie ; Zeng, Kai ; Feng, Mingtao ; Chang, Xiaojun ; Wang, Yaonan</creatorcontrib><description>In the electric power scene, fasteners play a pivotal role in securing and connecting electrical equipment, with small fastener detection (SFD) being crucial for ensuring operational stability. Despite the replacement of manual inspection methods by non-destructive techniques employing deep learning, these approaches often demand substantial computational resources and involve numerous parameters. While knowledge distillation (KD) can be a viable solution, existing KD methods may often fail to achieve satisfactory performance when dealing with small object presentation and little inter-class variability in SFD tasks. To alleviate this, we propose a Focal Multi-scale Shape-feature Distillation Network (FMSD) to achieve efficient and precise fastener detection in electric power scenarios. Specifically, we propose a novel Multi-Scale Shape-Aware Feature Aggregation module (MSFA) to augment the network's perception of object shape and scale during the KD process. Additionally, we propose a Contour-Guided Distillation (CGD) module to optimize the transfer of the extracted shape-sensitive knowledge between the teacher and student models. Through a series of experiments compared with existing state-of-the-art (SOTA) methods, our method demonstrates superior performance over existing SOTA techniques, both efficiently and effectively. Furthermore, validation on publicly available power scene datasets confirms the generalizability and adaptability of our proposed FMSD across various settings.</description><identifier>ISSN: 1051-8215</identifier><identifier>EISSN: 1558-2205</identifier><identifier>DOI: 10.1109/TCSVT.2024.3485548</identifier><identifier>CODEN: ITCTEM</identifier><language>eng</language><publisher>IEEE</publisher><subject>contour-guided distillation ; Detectors ; Fasteners ; Feature extraction ; Inspection ; knowledge distillation ; Knowledge engineering ; Object detection ; Pins ; Power systems ; Proposals ; shape-sensitive knowledge ; Small fastener detection ; Training</subject><ispartof>IEEE transactions on circuits and systems for video technology, 2024-10, p.1-1</ispartof><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><orcidid>0000-0003-2267-5415 ; 0000-0003-1737-0078 ; 0000-0002-1803-3148 ; 0000-0002-0519-6458 ; 0000-0002-7778-8807 ; 0000-0002-2745-1253</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10731891$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,776,780,792,27901,27902,54733</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10731891$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Yi, Junfei</creatorcontrib><creatorcontrib>Mao, Jianxu</creatorcontrib><creatorcontrib>Zhang, Hui</creatorcontrib><creatorcontrib>Li, Mingjie</creatorcontrib><creatorcontrib>Zeng, Kai</creatorcontrib><creatorcontrib>Feng, Mingtao</creatorcontrib><creatorcontrib>Chang, Xiaojun</creatorcontrib><creatorcontrib>Wang, Yaonan</creatorcontrib><title>FMSD: Focal Multi-scale Shape-feature Distillation Network for Small Fasteners Detection in Electric Power Scene</title><title>IEEE transactions on circuits and systems for video technology</title><addtitle>TCSVT</addtitle><description>In the electric power scene, fasteners play a pivotal role in securing and connecting electrical equipment, with small fastener detection (SFD) being crucial for ensuring operational stability. Despite the replacement of manual inspection methods by non-destructive techniques employing deep learning, these approaches often demand substantial computational resources and involve numerous parameters. While knowledge distillation (KD) can be a viable solution, existing KD methods may often fail to achieve satisfactory performance when dealing with small object presentation and little inter-class variability in SFD tasks. To alleviate this, we propose a Focal Multi-scale Shape-feature Distillation Network (FMSD) to achieve efficient and precise fastener detection in electric power scenarios. Specifically, we propose a novel Multi-Scale Shape-Aware Feature Aggregation module (MSFA) to augment the network's perception of object shape and scale during the KD process. Additionally, we propose a Contour-Guided Distillation (CGD) module to optimize the transfer of the extracted shape-sensitive knowledge between the teacher and student models. Through a series of experiments compared with existing state-of-the-art (SOTA) methods, our method demonstrates superior performance over existing SOTA techniques, both efficiently and effectively. Furthermore, validation on publicly available power scene datasets confirms the generalizability and adaptability of our proposed FMSD across various settings.</description><subject>contour-guided distillation</subject><subject>Detectors</subject><subject>Fasteners</subject><subject>Feature extraction</subject><subject>Inspection</subject><subject>knowledge distillation</subject><subject>Knowledge engineering</subject><subject>Object detection</subject><subject>Pins</subject><subject>Power systems</subject><subject>Proposals</subject><subject>shape-sensitive knowledge</subject><subject>Small fastener detection</subject><subject>Training</subject><issn>1051-8215</issn><issn>1558-2205</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNpNkMFKAzEURYMoWKs_IC7yA6lJJpnMuJO2o0KrwgxuhzTzgtG0U5KU4t87tV24endxz4V3ELpldMIYLe-baf3RTDjlYpKJQkpRnKERk7IgnFN5PmQqGSk4k5foKsYvSpkohBqhbbWsZw-46o32eLnzyZE4RMD1p94CsaDTLgCeuZic9zq5foNfIe378I1tH3C91t7jSscEGwgRzyCB-Wu5DZ77IQdn8Hu_h6Frhs41urDaR7g53TFqqnkzfSaLt6eX6eOCmFwwojrelYrZzuZ53ukyA21ykwlqMzBSQWHLjq-MVdQqpWDFjBZCDf9SKyy1Nhsjfpw1oY8xgG23wa11-GkZbQ_K2j9l7UFZe1I2QHdHyAHAP0BlrChZ9gsQJ2on</recordid><startdate>20241022</startdate><enddate>20241022</enddate><creator>Yi, Junfei</creator><creator>Mao, Jianxu</creator><creator>Zhang, Hui</creator><creator>Li, Mingjie</creator><creator>Zeng, Kai</creator><creator>Feng, Mingtao</creator><creator>Chang, Xiaojun</creator><creator>Wang, Yaonan</creator><general>IEEE</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope><orcidid>https://orcid.org/0000-0003-2267-5415</orcidid><orcidid>https://orcid.org/0000-0003-1737-0078</orcidid><orcidid>https://orcid.org/0000-0002-1803-3148</orcidid><orcidid>https://orcid.org/0000-0002-0519-6458</orcidid><orcidid>https://orcid.org/0000-0002-7778-8807</orcidid><orcidid>https://orcid.org/0000-0002-2745-1253</orcidid></search><sort><creationdate>20241022</creationdate><title>FMSD: Focal Multi-scale Shape-feature Distillation Network for Small Fasteners Detection in Electric Power Scene</title><author>Yi, Junfei ; Mao, Jianxu ; Zhang, Hui ; Li, Mingjie ; Zeng, Kai ; Feng, Mingtao ; Chang, Xiaojun ; Wang, Yaonan</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c641-7d2d971fdf666da93eac6c340f3ec57e8f9d2bcf70f777eb1ca4475580f4f0ff3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>contour-guided distillation</topic><topic>Detectors</topic><topic>Fasteners</topic><topic>Feature extraction</topic><topic>Inspection</topic><topic>knowledge distillation</topic><topic>Knowledge engineering</topic><topic>Object detection</topic><topic>Pins</topic><topic>Power systems</topic><topic>Proposals</topic><topic>shape-sensitive knowledge</topic><topic>Small fastener detection</topic><topic>Training</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Yi, Junfei</creatorcontrib><creatorcontrib>Mao, Jianxu</creatorcontrib><creatorcontrib>Zhang, Hui</creatorcontrib><creatorcontrib>Li, Mingjie</creatorcontrib><creatorcontrib>Zeng, Kai</creatorcontrib><creatorcontrib>Feng, Mingtao</creatorcontrib><creatorcontrib>Chang, Xiaojun</creatorcontrib><creatorcontrib>Wang, Yaonan</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998–Present</collection><collection>IEEE</collection><collection>CrossRef</collection><jtitle>IEEE transactions on circuits and systems for video technology</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Yi, Junfei</au><au>Mao, Jianxu</au><au>Zhang, Hui</au><au>Li, Mingjie</au><au>Zeng, Kai</au><au>Feng, Mingtao</au><au>Chang, Xiaojun</au><au>Wang, Yaonan</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>FMSD: Focal Multi-scale Shape-feature Distillation Network for Small Fasteners Detection in Electric Power Scene</atitle><jtitle>IEEE transactions on circuits and systems for video technology</jtitle><stitle>TCSVT</stitle><date>2024-10-22</date><risdate>2024</risdate><spage>1</spage><epage>1</epage><pages>1-1</pages><issn>1051-8215</issn><eissn>1558-2205</eissn><coden>ITCTEM</coden><abstract>In the electric power scene, fasteners play a pivotal role in securing and connecting electrical equipment, with small fastener detection (SFD) being crucial for ensuring operational stability. Despite the replacement of manual inspection methods by non-destructive techniques employing deep learning, these approaches often demand substantial computational resources and involve numerous parameters. While knowledge distillation (KD) can be a viable solution, existing KD methods may often fail to achieve satisfactory performance when dealing with small object presentation and little inter-class variability in SFD tasks. To alleviate this, we propose a Focal Multi-scale Shape-feature Distillation Network (FMSD) to achieve efficient and precise fastener detection in electric power scenarios. Specifically, we propose a novel Multi-Scale Shape-Aware Feature Aggregation module (MSFA) to augment the network's perception of object shape and scale during the KD process. Additionally, we propose a Contour-Guided Distillation (CGD) module to optimize the transfer of the extracted shape-sensitive knowledge between the teacher and student models. Through a series of experiments compared with existing state-of-the-art (SOTA) methods, our method demonstrates superior performance over existing SOTA techniques, both efficiently and effectively. Furthermore, validation on publicly available power scene datasets confirms the generalizability and adaptability of our proposed FMSD across various settings.</abstract><pub>IEEE</pub><doi>10.1109/TCSVT.2024.3485548</doi><tpages>1</tpages><orcidid>https://orcid.org/0000-0003-2267-5415</orcidid><orcidid>https://orcid.org/0000-0003-1737-0078</orcidid><orcidid>https://orcid.org/0000-0002-1803-3148</orcidid><orcidid>https://orcid.org/0000-0002-0519-6458</orcidid><orcidid>https://orcid.org/0000-0002-7778-8807</orcidid><orcidid>https://orcid.org/0000-0002-2745-1253</orcidid></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | ISSN: 1051-8215 |
ispartof | IEEE transactions on circuits and systems for video technology, 2024-10, p.1-1 |
issn | 1051-8215 1558-2205 |
language | eng |
recordid | cdi_crossref_primary_10_1109_TCSVT_2024_3485548 |
source | IEEE |
subjects | contour-guided distillation Detectors Fasteners Feature extraction Inspection knowledge distillation Knowledge engineering Object detection Pins Power systems Proposals shape-sensitive knowledge Small fastener detection Training |
title | FMSD: Focal Multi-scale Shape-feature Distillation Network for Small Fasteners Detection in Electric Power Scene |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-01T12%3A07%3A20IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-crossref_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=FMSD:%20Focal%20Multi-scale%20Shape-feature%20Distillation%20Network%20for%20Small%20Fasteners%20Detection%20in%20Electric%20Power%20Scene&rft.jtitle=IEEE%20transactions%20on%20circuits%20and%20systems%20for%20video%20technology&rft.au=Yi,%20Junfei&rft.date=2024-10-22&rft.spage=1&rft.epage=1&rft.pages=1-1&rft.issn=1051-8215&rft.eissn=1558-2205&rft.coden=ITCTEM&rft_id=info:doi/10.1109/TCSVT.2024.3485548&rft_dat=%3Ccrossref_RIE%3E10_1109_TCSVT_2024_3485548%3C/crossref_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=10731891&rfr_iscdi=true |