Loading…
Self-Supervised Industrial Image Generation for Defect Recognition Under Limited Data
Deep learning-based industrial image inspection has achieved tremendous success in recent years, while the task of learning a model from a few defect images remains unexplored. The most popular method is to directly generate defect images to augment the limited dataset before training the recognitio...
Saved in:
| Main Author: | |
|---|---|
| Format: | Conference Proceeding |
| Language: | English |
| Subjects: | |
| Online Access: | Request full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| cited_by | |
|---|---|
| cites | |
| container_end_page | 7 |
| container_issue | |
| container_start_page | 1 |
| container_title | |
| container_volume | |
| creator | Lee, Younkwan |
| description | Deep learning-based industrial image inspection has achieved tremendous success in recent years, while the task of learning a model from a few defect images remains unexplored. The most popular method is to directly generate defect images to augment the limited dataset before training the recognition model. To address this issue, we propose a new framework called REG-Net that combines the generation module and classification module in an end-to-end manner, with few-shot defect image generation as assistance. Specifically, we design a two-branch module with attention fusion to directly combine normal and defect features. This reduces the negative impact on the classification module when the generation module performs poorly. REGNet improves further classification performance through recognition-friendly defect generation. Extensive experiments on the MVTec-AD benchmark show that effectiveness compared to the recent state-of-the-art methods. |
| doi_str_mv | 10.1109/AVSS61716.2024.10672565 |
| format | conference_proceeding |
| fullrecord | <record><control><sourceid>ieee_CHZPO</sourceid><recordid>TN_cdi_ieee_primary_10672565</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>10672565</ieee_id><sourcerecordid>10672565</sourcerecordid><originalsourceid>FETCH-LOGICAL-i106t-cdf11c67b63b235758b5ecf8028819cf7ec251125b22687120cf992b1e876a683</originalsourceid><addsrcrecordid>eNo1kNtKAzEURaMgWOr8gWB-YGrOyeT2WFptCwXBsb6WTOakRNppyUwF_97i5WnBflhsFmMPICYAwj1O3-tagwE9QYHVBIQ2qLS6YoUzzkolpKnQqms2Ql3JUiPIW1b0_YcQQoJ11sGIbWrax7I-nyh_pp5avuracz_k5Pd8dfA74gvqKPshHTsej5nPKVIY-CuF465LP_OmaynzdTqk4SKY-8HfsZvo9z0VfxyzzfPT22xZrl8Wq9l0XabL26EMbQQI2jRaNiiVUbZRFKIVaC24EA0FVACoGkRtDaAI0TlsgKzRXls5Zve_3kRE21NOB5-_tv8l5DeRxVJb</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>Self-Supervised Industrial Image Generation for Defect Recognition Under Limited Data</title><source>IEEE Xplore All Conference Series</source><creator>Lee, Younkwan</creator><creatorcontrib>Lee, Younkwan</creatorcontrib><description>Deep learning-based industrial image inspection has achieved tremendous success in recent years, while the task of learning a model from a few defect images remains unexplored. The most popular method is to directly generate defect images to augment the limited dataset before training the recognition model. To address this issue, we propose a new framework called REG-Net that combines the generation module and classification module in an end-to-end manner, with few-shot defect image generation as assistance. Specifically, we design a two-branch module with attention fusion to directly combine normal and defect features. This reduces the negative impact on the classification module when the generation module performs poorly. REGNet improves further classification performance through recognition-friendly defect generation. Extensive experiments on the MVTec-AD benchmark show that effectiveness compared to the recent state-of-the-art methods.</description><identifier>EISSN: 2643-6213</identifier><identifier>EISBN: 9798350374285</identifier><identifier>DOI: 10.1109/AVSS61716.2024.10672565</identifier><language>eng</language><publisher>IEEE</publisher><subject>Accuracy ; Benchmark testing ; Image recognition ; Image synthesis ; Inspection ; Surveillance ; Training</subject><ispartof>2024 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2024, p.1-7</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/10672565$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,776,780,785,786,27902,54530,54907</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/10672565$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Lee, Younkwan</creatorcontrib><title>Self-Supervised Industrial Image Generation for Defect Recognition Under Limited Data</title><title>2024 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)</title><addtitle>AVSS</addtitle><description>Deep learning-based industrial image inspection has achieved tremendous success in recent years, while the task of learning a model from a few defect images remains unexplored. The most popular method is to directly generate defect images to augment the limited dataset before training the recognition model. To address this issue, we propose a new framework called REG-Net that combines the generation module and classification module in an end-to-end manner, with few-shot defect image generation as assistance. Specifically, we design a two-branch module with attention fusion to directly combine normal and defect features. This reduces the negative impact on the classification module when the generation module performs poorly. REGNet improves further classification performance through recognition-friendly defect generation. Extensive experiments on the MVTec-AD benchmark show that effectiveness compared to the recent state-of-the-art methods.</description><subject>Accuracy</subject><subject>Benchmark testing</subject><subject>Image recognition</subject><subject>Image synthesis</subject><subject>Inspection</subject><subject>Surveillance</subject><subject>Training</subject><issn>2643-6213</issn><isbn>9798350374285</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2024</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><recordid>eNo1kNtKAzEURaMgWOr8gWB-YGrOyeT2WFptCwXBsb6WTOakRNppyUwF_97i5WnBflhsFmMPICYAwj1O3-tagwE9QYHVBIQ2qLS6YoUzzkolpKnQqms2Ql3JUiPIW1b0_YcQQoJ11sGIbWrax7I-nyh_pp5avuracz_k5Pd8dfA74gvqKPshHTsej5nPKVIY-CuF465LP_OmaynzdTqk4SKY-8HfsZvo9z0VfxyzzfPT22xZrl8Wq9l0XabL26EMbQQI2jRaNiiVUbZRFKIVaC24EA0FVACoGkRtDaAI0TlsgKzRXls5Zve_3kRE21NOB5-_tv8l5DeRxVJb</recordid><startdate>20240715</startdate><enddate>20240715</enddate><creator>Lee, Younkwan</creator><general>IEEE</general><scope>6IE</scope><scope>6IL</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIL</scope></search><sort><creationdate>20240715</creationdate><title>Self-Supervised Industrial Image Generation for Defect Recognition Under Limited Data</title><author>Lee, Younkwan</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i106t-cdf11c67b63b235758b5ecf8028819cf7ec251125b22687120cf992b1e876a683</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Accuracy</topic><topic>Benchmark testing</topic><topic>Image recognition</topic><topic>Image synthesis</topic><topic>Inspection</topic><topic>Surveillance</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Lee, Younkwan</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan All Online (POP All Online) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Electronic Library (IEL)</collection><collection>IEEE Proceedings Order Plans (POP All) 1998-Present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Lee, Younkwan</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>Self-Supervised Industrial Image Generation for Defect Recognition Under Limited Data</atitle><btitle>2024 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)</btitle><stitle>AVSS</stitle><date>2024-07-15</date><risdate>2024</risdate><spage>1</spage><epage>7</epage><pages>1-7</pages><eissn>2643-6213</eissn><eisbn>9798350374285</eisbn><abstract>Deep learning-based industrial image inspection has achieved tremendous success in recent years, while the task of learning a model from a few defect images remains unexplored. The most popular method is to directly generate defect images to augment the limited dataset before training the recognition model. To address this issue, we propose a new framework called REG-Net that combines the generation module and classification module in an end-to-end manner, with few-shot defect image generation as assistance. Specifically, we design a two-branch module with attention fusion to directly combine normal and defect features. This reduces the negative impact on the classification module when the generation module performs poorly. REGNet improves further classification performance through recognition-friendly defect generation. Extensive experiments on the MVTec-AD benchmark show that effectiveness compared to the recent state-of-the-art methods.</abstract><pub>IEEE</pub><doi>10.1109/AVSS61716.2024.10672565</doi><tpages>7</tpages></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | EISSN: 2643-6213 |
| ispartof | 2024 IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2024, p.1-7 |
| issn | 2643-6213 |
| language | eng |
| recordid | cdi_ieee_primary_10672565 |
| source | IEEE Xplore All Conference Series |
| subjects | Accuracy Benchmark testing Image recognition Image synthesis Inspection Surveillance Training |
| title | Self-Supervised Industrial Image Generation for Defect Recognition Under Limited Data |
| url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-03T00%3A30%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=Self-Supervised%20Industrial%20Image%20Generation%20for%20Defect%20Recognition%20Under%20Limited%20Data&rft.btitle=2024%20IEEE%20International%20Conference%20on%20Advanced%20Video%20and%20Signal%20Based%20Surveillance%20(AVSS)&rft.au=Lee,%20Younkwan&rft.date=2024-07-15&rft.spage=1&rft.epage=7&rft.pages=1-7&rft.eissn=2643-6213&rft_id=info:doi/10.1109/AVSS61716.2024.10672565&rft.eisbn=9798350374285&rft_dat=%3Cieee_CHZPO%3E10672565%3C/ieee_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-i106t-cdf11c67b63b235758b5ecf8028819cf7ec251125b22687120cf992b1e876a683%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=10672565&rfr_iscdi=true |