AI can evolve without labels: self-evolving vision transformer for chest X-ray diagnosis through knowledge distillation
Although deep learning-based computer-aided diagnosis systems have recently achieved expert-level performance, developing a robust deep learning model requires large, high-quality data with manual annotation, which is expensive to obtain. This situation poses the problem that the chest x-rays collec...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2022-02 |
---|---|
Hauptverfasser: | , , , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Park, Sangjoon Kim, Gwanghyun Oh, Yujin Seo, Joon Beom Lee, Sang Min Kim, Jin Hwan Moon, Sungjun Lim, Jae-Kwang Park, Chang Min Ye, Jong Chul |
description | Although deep learning-based computer-aided diagnosis systems have recently achieved expert-level performance, developing a robust deep learning model requires large, high-quality data with manual annotation, which is expensive to obtain. This situation poses the problem that the chest x-rays collected annually in hospitals cannot be used due to the lack of manual labeling by experts, especially in deprived areas. To address this, here we present a novel deep learning framework that uses knowledge distillation through self-supervised learning and self-training, which shows that the performance of the original model trained with a small number of labels can be gradually improved with more unlabeled data. Experimental results show that the proposed framework maintains impressive robustness against a real-world environment and has general applicability to several diagnostic tasks such as tuberculosis, pneumothorax, and COVID-19. Notably, we demonstrated that our model performs even better than those trained with the same amount of labeled data. The proposed framework has a great potential for medical imaging, where plenty of data is accumulated every year, but ground truth annotations are expensive to obtain. |
doi_str_mv | 10.48550/arxiv.2202.06431 |
format | Article |
fullrecord | <record><control><sourceid>proquest_arxiv</sourceid><recordid>TN_cdi_arxiv_primary_2202_06431</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2628908563</sourcerecordid><originalsourceid>FETCH-LOGICAL-a951-c2db2d19c18f1af345e4cc13d5d918bfbf97470ddceb0ffc7c7555e8a8c154083</originalsourceid><addsrcrecordid>eNotkDtPwzAAhC0kJKrSH8CEJeYUPxOHrap4VKrE0oEtcvxIXNy42ElK_z2hMN1wp9PdB8AdRksmOEePMn67cUkIIkuUM4qvwIxQijPBCLkBi5T2CCGSF4RzOgOn1QYq2UEzBj8aeHJ9G4Yeelkbn55gMt5mF891DRxdcqGDfZRdsiEeTISTQNWa1MOPLMoz1E42XUguwb6NYWha-NmFkze6MZOXeue97KeSW3BtpU9m8a9zsHt53q3fsu3762a92may5DhTRNdE41JhYbG0lHHDlMJUc11iUdvalgUrkNbK1MhaVaiCc26EFApzhgSdg_u_2guV6hjdQcZz9UunutCZEg9_iWMMX8N0pNqHIXbTporkRJRI8JzSH9BBauM</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2628908563</pqid></control><display><type>article</type><title>AI can evolve without labels: self-evolving vision transformer for chest X-ray diagnosis through knowledge distillation</title><source>arXiv.org</source><source>Free E- Journals</source><creator>Park, Sangjoon ; Kim, Gwanghyun ; Oh, Yujin ; Seo, Joon Beom ; Lee, Sang Min ; Kim, Jin Hwan ; Moon, Sungjun ; Lim, Jae-Kwang ; Park, Chang Min ; Ye, Jong Chul</creator><creatorcontrib>Park, Sangjoon ; Kim, Gwanghyun ; Oh, Yujin ; Seo, Joon Beom ; Lee, Sang Min ; Kim, Jin Hwan ; Moon, Sungjun ; Lim, Jae-Kwang ; Park, Chang Min ; Ye, Jong Chul</creatorcontrib><description>Although deep learning-based computer-aided diagnosis systems have recently achieved expert-level performance, developing a robust deep learning model requires large, high-quality data with manual annotation, which is expensive to obtain. This situation poses the problem that the chest x-rays collected annually in hospitals cannot be used due to the lack of manual labeling by experts, especially in deprived areas. To address this, here we present a novel deep learning framework that uses knowledge distillation through self-supervised learning and self-training, which shows that the performance of the original model trained with a small number of labels can be gradually improved with more unlabeled data. Experimental results show that the proposed framework maintains impressive robustness against a real-world environment and has general applicability to several diagnostic tasks such as tuberculosis, pneumothorax, and COVID-19. Notably, we demonstrated that our model performs even better than those trained with the same amount of labeled data. The proposed framework has a great potential for medical imaging, where plenty of data is accumulated every year, but ground truth annotations are expensive to obtain.</description><identifier>EISSN: 2331-8422</identifier><identifier>DOI: 10.48550/arxiv.2202.06431</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Annotations ; CAI ; Chest ; Computer assisted instruction ; Computer Science - Computer Vision and Pattern Recognition ; Computer Science - Learning ; Deep learning ; Diagnosis ; Distillation ; Labels ; Medical imaging ; Pneumothorax ; Tuberculosis</subject><ispartof>arXiv.org, 2022-02</ispartof><rights>2022. This work is published under http://creativecommons.org/licenses/by-nc-nd/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><rights>http://creativecommons.org/licenses/by-nc-nd/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,780,881,27902</link.rule.ids><backlink>$$Uhttps://doi.org/10.1038/s41467-022-31514-x$$DView published paper (Access to full text may be restricted)$$Hfree_for_read</backlink><backlink>$$Uhttps://doi.org/10.48550/arXiv.2202.06431$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Park, Sangjoon</creatorcontrib><creatorcontrib>Kim, Gwanghyun</creatorcontrib><creatorcontrib>Oh, Yujin</creatorcontrib><creatorcontrib>Seo, Joon Beom</creatorcontrib><creatorcontrib>Lee, Sang Min</creatorcontrib><creatorcontrib>Kim, Jin Hwan</creatorcontrib><creatorcontrib>Moon, Sungjun</creatorcontrib><creatorcontrib>Lim, Jae-Kwang</creatorcontrib><creatorcontrib>Park, Chang Min</creatorcontrib><creatorcontrib>Ye, Jong Chul</creatorcontrib><title>AI can evolve without labels: self-evolving vision transformer for chest X-ray diagnosis through knowledge distillation</title><title>arXiv.org</title><description>Although deep learning-based computer-aided diagnosis systems have recently achieved expert-level performance, developing a robust deep learning model requires large, high-quality data with manual annotation, which is expensive to obtain. This situation poses the problem that the chest x-rays collected annually in hospitals cannot be used due to the lack of manual labeling by experts, especially in deprived areas. To address this, here we present a novel deep learning framework that uses knowledge distillation through self-supervised learning and self-training, which shows that the performance of the original model trained with a small number of labels can be gradually improved with more unlabeled data. Experimental results show that the proposed framework maintains impressive robustness against a real-world environment and has general applicability to several diagnostic tasks such as tuberculosis, pneumothorax, and COVID-19. Notably, we demonstrated that our model performs even better than those trained with the same amount of labeled data. The proposed framework has a great potential for medical imaging, where plenty of data is accumulated every year, but ground truth annotations are expensive to obtain.</description><subject>Annotations</subject><subject>CAI</subject><subject>Chest</subject><subject>Computer assisted instruction</subject><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Computer Science - Learning</subject><subject>Deep learning</subject><subject>Diagnosis</subject><subject>Distillation</subject><subject>Labels</subject><subject>Medical imaging</subject><subject>Pneumothorax</subject><subject>Tuberculosis</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>BENPR</sourceid><sourceid>GOX</sourceid><recordid>eNotkDtPwzAAhC0kJKrSH8CEJeYUPxOHrap4VKrE0oEtcvxIXNy42ElK_z2hMN1wp9PdB8AdRksmOEePMn67cUkIIkuUM4qvwIxQijPBCLkBi5T2CCGSF4RzOgOn1QYq2UEzBj8aeHJ9G4Yeelkbn55gMt5mF891DRxdcqGDfZRdsiEeTISTQNWa1MOPLMoz1E42XUguwb6NYWha-NmFkze6MZOXeue97KeSW3BtpU9m8a9zsHt53q3fsu3762a92may5DhTRNdE41JhYbG0lHHDlMJUc11iUdvalgUrkNbK1MhaVaiCc26EFApzhgSdg_u_2guV6hjdQcZz9UunutCZEg9_iWMMX8N0pNqHIXbTporkRJRI8JzSH9BBauM</recordid><startdate>20220213</startdate><enddate>20220213</enddate><creator>Park, Sangjoon</creator><creator>Kim, Gwanghyun</creator><creator>Oh, Yujin</creator><creator>Seo, Joon Beom</creator><creator>Lee, Sang Min</creator><creator>Kim, Jin Hwan</creator><creator>Moon, Sungjun</creator><creator>Lim, Jae-Kwang</creator><creator>Park, Chang Min</creator><creator>Ye, Jong Chul</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>COVID</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20220213</creationdate><title>AI can evolve without labels: self-evolving vision transformer for chest X-ray diagnosis through knowledge distillation</title><author>Park, Sangjoon ; Kim, Gwanghyun ; Oh, Yujin ; Seo, Joon Beom ; Lee, Sang Min ; Kim, Jin Hwan ; Moon, Sungjun ; Lim, Jae-Kwang ; Park, Chang Min ; Ye, Jong Chul</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a951-c2db2d19c18f1af345e4cc13d5d918bfbf97470ddceb0ffc7c7555e8a8c154083</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Annotations</topic><topic>CAI</topic><topic>Chest</topic><topic>Computer assisted instruction</topic><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Computer Science - Learning</topic><topic>Deep learning</topic><topic>Diagnosis</topic><topic>Distillation</topic><topic>Labels</topic><topic>Medical imaging</topic><topic>Pneumothorax</topic><topic>Tuberculosis</topic><toplevel>online_resources</toplevel><creatorcontrib>Park, Sangjoon</creatorcontrib><creatorcontrib>Kim, Gwanghyun</creatorcontrib><creatorcontrib>Oh, Yujin</creatorcontrib><creatorcontrib>Seo, Joon Beom</creatorcontrib><creatorcontrib>Lee, Sang Min</creatorcontrib><creatorcontrib>Kim, Jin Hwan</creatorcontrib><creatorcontrib>Moon, Sungjun</creatorcontrib><creatorcontrib>Lim, Jae-Kwang</creatorcontrib><creatorcontrib>Park, Chang Min</creatorcontrib><creatorcontrib>Ye, Jong Chul</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection (ProQuest)</collection><collection>ProQuest One Community College</collection><collection>Coronavirus Research Database</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content (ProQuest)</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection><collection>arXiv Computer Science</collection><collection>arXiv.org</collection><jtitle>arXiv.org</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Park, Sangjoon</au><au>Kim, Gwanghyun</au><au>Oh, Yujin</au><au>Seo, Joon Beom</au><au>Lee, Sang Min</au><au>Kim, Jin Hwan</au><au>Moon, Sungjun</au><au>Lim, Jae-Kwang</au><au>Park, Chang Min</au><au>Ye, Jong Chul</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>AI can evolve without labels: self-evolving vision transformer for chest X-ray diagnosis through knowledge distillation</atitle><jtitle>arXiv.org</jtitle><date>2022-02-13</date><risdate>2022</risdate><eissn>2331-8422</eissn><abstract>Although deep learning-based computer-aided diagnosis systems have recently achieved expert-level performance, developing a robust deep learning model requires large, high-quality data with manual annotation, which is expensive to obtain. This situation poses the problem that the chest x-rays collected annually in hospitals cannot be used due to the lack of manual labeling by experts, especially in deprived areas. To address this, here we present a novel deep learning framework that uses knowledge distillation through self-supervised learning and self-training, which shows that the performance of the original model trained with a small number of labels can be gradually improved with more unlabeled data. Experimental results show that the proposed framework maintains impressive robustness against a real-world environment and has general applicability to several diagnostic tasks such as tuberculosis, pneumothorax, and COVID-19. Notably, we demonstrated that our model performs even better than those trained with the same amount of labeled data. The proposed framework has a great potential for medical imaging, where plenty of data is accumulated every year, but ground truth annotations are expensive to obtain.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><doi>10.48550/arxiv.2202.06431</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2022-02 |
issn | 2331-8422 |
language | eng |
recordid | cdi_arxiv_primary_2202_06431 |
source | arXiv.org; Free E- Journals |
subjects | Annotations CAI Chest Computer assisted instruction Computer Science - Computer Vision and Pattern Recognition Computer Science - Learning Deep learning Diagnosis Distillation Labels Medical imaging Pneumothorax Tuberculosis |
title | AI can evolve without labels: self-evolving vision transformer for chest X-ray diagnosis through knowledge distillation |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-07T06%3A49%3A16IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_arxiv&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=AI%20can%20evolve%20without%20labels:%20self-evolving%20vision%20transformer%20for%20chest%20X-ray%20diagnosis%20through%20knowledge%20distillation&rft.jtitle=arXiv.org&rft.au=Park,%20Sangjoon&rft.date=2022-02-13&rft.eissn=2331-8422&rft_id=info:doi/10.48550/arxiv.2202.06431&rft_dat=%3Cproquest_arxiv%3E2628908563%3C/proquest_arxiv%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2628908563&rft_id=info:pmid/&rfr_iscdi=true |