ESRL: efficient similarity representation learning for deepfake detection

For Deepfake detection, many existing works use the cross-entropy loss to enforce the classifier network to learn the mapping relationship from the RGB domain to the class domain, lacking an explicit constraint to guide the feature extraction network to learn discriminative features from an input im...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Multimedia tools and applications 2024-02, Vol.83 (31), p.76991-77007
Hauptverfasser: Wang, Feng, Zhang, Dengyong, Guo, Zhiqing, Wang, Dewang, Yang, Gaobo
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 77007
container_issue 31
container_start_page 76991
container_title Multimedia tools and applications
container_volume 83
creator Wang, Feng
Zhang, Dengyong
Guo, Zhiqing
Wang, Dewang
Yang, Gaobo
description For Deepfake detection, many existing works use the cross-entropy loss to enforce the classifier network to learn the mapping relationship from the RGB domain to the class domain, lacking an explicit constraint to guide the feature extraction network to learn discriminative features from an input image. This constrains the feature representation capability to expose deepfake. In this work, we analyze the feature extraction network in terms of both difference and similarity capabilities and propose a new constraint called similarity loss (SL) to improve the detection performance of the convolutional neural network (CNN) based detector. Moreover, according to the experimental results of the SL on data augmentation effectiveness, we propose a simple yet efficient framework, which is called as efficient similarity representation learning (ESRL), for deepfake detection. Extensive experiments on three public datasets (namely FF++, DFDC, and Celeb-DF) show that the feature extraction network trained with the help of SL can map forged faces and real faces to different feature embedding and map the same type of forged faces to similar feature embedding.
doi_str_mv 10.1007/s11042-024-18447-x
format Article
fullrecord <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_journals_3100357016</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3100357016</sourcerecordid><originalsourceid>FETCH-LOGICAL-c270t-305717b5f56820b7fcf309762c11baa2f5d766745c817d0f4f5131ba854ddb713</originalsourceid><addsrcrecordid>eNp9kE1LAzEQhoMoWKt_wNOC59VMPnZWb1KqFgqCH-eQzSYldbu7Jlto_72pK-jJ07zMvO8M8xByCfQaKMWbCEAFyykTOZRCYL47IhOQyHNEBsd_9Ck5i3FNKRSSiQlZzF9flneZdc4bb9shi37jGx38sM-C7YONqakH37VZY3VofbvKXBey2tre6Q-bxGDNYX5OTpxuor34qVPy_jB_mz3ly-fHxex-mRuGdMg5lQhYSSeLktEKnXGc3mLBDEClNXOyxqJAIU0JWFMnnASeJqUUdV0h8Cm5Gvf2ofvc2jiodbcNbTqpeGLBJabfkouNLhO6GIN1qg9-o8NeAVUHZGpEphIy9Y1M7VKIj6GYzO3Kht_V_6S-AKdtbsY</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3100357016</pqid></control><display><type>article</type><title>ESRL: efficient similarity representation learning for deepfake detection</title><source>SpringerLink Journals - AutoHoldings</source><creator>Wang, Feng ; Zhang, Dengyong ; Guo, Zhiqing ; Wang, Dewang ; Yang, Gaobo</creator><creatorcontrib>Wang, Feng ; Zhang, Dengyong ; Guo, Zhiqing ; Wang, Dewang ; Yang, Gaobo</creatorcontrib><description>For Deepfake detection, many existing works use the cross-entropy loss to enforce the classifier network to learn the mapping relationship from the RGB domain to the class domain, lacking an explicit constraint to guide the feature extraction network to learn discriminative features from an input image. This constrains the feature representation capability to expose deepfake. In this work, we analyze the feature extraction network in terms of both difference and similarity capabilities and propose a new constraint called similarity loss (SL) to improve the detection performance of the convolutional neural network (CNN) based detector. Moreover, according to the experimental results of the SL on data augmentation effectiveness, we propose a simple yet efficient framework, which is called as efficient similarity representation learning (ESRL), for deepfake detection. Extensive experiments on three public datasets (namely FF++, DFDC, and Celeb-DF) show that the feature extraction network trained with the help of SL can map forged faces and real faces to different feature embedding and map the same type of forged faces to similar feature embedding.</description><identifier>ISSN: 1573-7721</identifier><identifier>ISSN: 1380-7501</identifier><identifier>EISSN: 1573-7721</identifier><identifier>DOI: 10.1007/s11042-024-18447-x</identifier><language>eng</language><publisher>New York: Springer US</publisher><subject>Artificial neural networks ; Computer Communication Networks ; Computer Science ; Constraints ; Data augmentation ; Data Structures and Information Theory ; Deception ; Deepfake ; Embedding ; Feature extraction ; Machine learning ; Multimedia Information Systems ; Representations ; Similarity ; Special Purpose and Application-Based Systems ; Track 6: Computer Vision for Multimedia Applications</subject><ispartof>Multimedia tools and applications, 2024-02, Vol.83 (31), p.76991-77007</ispartof><rights>The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2024. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><cites>FETCH-LOGICAL-c270t-305717b5f56820b7fcf309762c11baa2f5d766745c817d0f4f5131ba854ddb713</cites><orcidid>0000-0003-2734-659X</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/s11042-024-18447-x$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/s11042-024-18447-x$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>314,780,784,27924,27925,41488,42557,51319</link.rule.ids></links><search><creatorcontrib>Wang, Feng</creatorcontrib><creatorcontrib>Zhang, Dengyong</creatorcontrib><creatorcontrib>Guo, Zhiqing</creatorcontrib><creatorcontrib>Wang, Dewang</creatorcontrib><creatorcontrib>Yang, Gaobo</creatorcontrib><title>ESRL: efficient similarity representation learning for deepfake detection</title><title>Multimedia tools and applications</title><addtitle>Multimed Tools Appl</addtitle><description>For Deepfake detection, many existing works use the cross-entropy loss to enforce the classifier network to learn the mapping relationship from the RGB domain to the class domain, lacking an explicit constraint to guide the feature extraction network to learn discriminative features from an input image. This constrains the feature representation capability to expose deepfake. In this work, we analyze the feature extraction network in terms of both difference and similarity capabilities and propose a new constraint called similarity loss (SL) to improve the detection performance of the convolutional neural network (CNN) based detector. Moreover, according to the experimental results of the SL on data augmentation effectiveness, we propose a simple yet efficient framework, which is called as efficient similarity representation learning (ESRL), for deepfake detection. Extensive experiments on three public datasets (namely FF++, DFDC, and Celeb-DF) show that the feature extraction network trained with the help of SL can map forged faces and real faces to different feature embedding and map the same type of forged faces to similar feature embedding.</description><subject>Artificial neural networks</subject><subject>Computer Communication Networks</subject><subject>Computer Science</subject><subject>Constraints</subject><subject>Data augmentation</subject><subject>Data Structures and Information Theory</subject><subject>Deception</subject><subject>Deepfake</subject><subject>Embedding</subject><subject>Feature extraction</subject><subject>Machine learning</subject><subject>Multimedia Information Systems</subject><subject>Representations</subject><subject>Similarity</subject><subject>Special Purpose and Application-Based Systems</subject><subject>Track 6: Computer Vision for Multimedia Applications</subject><issn>1573-7721</issn><issn>1380-7501</issn><issn>1573-7721</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><recordid>eNp9kE1LAzEQhoMoWKt_wNOC59VMPnZWb1KqFgqCH-eQzSYldbu7Jlto_72pK-jJ07zMvO8M8xByCfQaKMWbCEAFyykTOZRCYL47IhOQyHNEBsd_9Ck5i3FNKRSSiQlZzF9flneZdc4bb9shi37jGx38sM-C7YONqakH37VZY3VofbvKXBey2tre6Q-bxGDNYX5OTpxuor34qVPy_jB_mz3ly-fHxex-mRuGdMg5lQhYSSeLktEKnXGc3mLBDEClNXOyxqJAIU0JWFMnnASeJqUUdV0h8Cm5Gvf2ofvc2jiodbcNbTqpeGLBJabfkouNLhO6GIN1qg9-o8NeAVUHZGpEphIy9Y1M7VKIj6GYzO3Kht_V_6S-AKdtbsY</recordid><startdate>20240221</startdate><enddate>20240221</enddate><creator>Wang, Feng</creator><creator>Zhang, Dengyong</creator><creator>Guo, Zhiqing</creator><creator>Wang, Dewang</creator><creator>Yang, Gaobo</creator><general>Springer US</general><general>Springer Nature B.V</general><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><orcidid>https://orcid.org/0000-0003-2734-659X</orcidid></search><sort><creationdate>20240221</creationdate><title>ESRL: efficient similarity representation learning for deepfake detection</title><author>Wang, Feng ; Zhang, Dengyong ; Guo, Zhiqing ; Wang, Dewang ; Yang, Gaobo</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c270t-305717b5f56820b7fcf309762c11baa2f5d766745c817d0f4f5131ba854ddb713</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Artificial neural networks</topic><topic>Computer Communication Networks</topic><topic>Computer Science</topic><topic>Constraints</topic><topic>Data augmentation</topic><topic>Data Structures and Information Theory</topic><topic>Deception</topic><topic>Deepfake</topic><topic>Embedding</topic><topic>Feature extraction</topic><topic>Machine learning</topic><topic>Multimedia Information Systems</topic><topic>Representations</topic><topic>Similarity</topic><topic>Special Purpose and Application-Based Systems</topic><topic>Track 6: Computer Vision for Multimedia Applications</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Wang, Feng</creatorcontrib><creatorcontrib>Zhang, Dengyong</creatorcontrib><creatorcontrib>Guo, Zhiqing</creatorcontrib><creatorcontrib>Wang, Dewang</creatorcontrib><creatorcontrib>Yang, Gaobo</creatorcontrib><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><jtitle>Multimedia tools and applications</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Wang, Feng</au><au>Zhang, Dengyong</au><au>Guo, Zhiqing</au><au>Wang, Dewang</au><au>Yang, Gaobo</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>ESRL: efficient similarity representation learning for deepfake detection</atitle><jtitle>Multimedia tools and applications</jtitle><stitle>Multimed Tools Appl</stitle><date>2024-02-21</date><risdate>2024</risdate><volume>83</volume><issue>31</issue><spage>76991</spage><epage>77007</epage><pages>76991-77007</pages><issn>1573-7721</issn><issn>1380-7501</issn><eissn>1573-7721</eissn><abstract>For Deepfake detection, many existing works use the cross-entropy loss to enforce the classifier network to learn the mapping relationship from the RGB domain to the class domain, lacking an explicit constraint to guide the feature extraction network to learn discriminative features from an input image. This constrains the feature representation capability to expose deepfake. In this work, we analyze the feature extraction network in terms of both difference and similarity capabilities and propose a new constraint called similarity loss (SL) to improve the detection performance of the convolutional neural network (CNN) based detector. Moreover, according to the experimental results of the SL on data augmentation effectiveness, we propose a simple yet efficient framework, which is called as efficient similarity representation learning (ESRL), for deepfake detection. Extensive experiments on three public datasets (namely FF++, DFDC, and Celeb-DF) show that the feature extraction network trained with the help of SL can map forged faces and real faces to different feature embedding and map the same type of forged faces to similar feature embedding.</abstract><cop>New York</cop><pub>Springer US</pub><doi>10.1007/s11042-024-18447-x</doi><tpages>17</tpages><orcidid>https://orcid.org/0000-0003-2734-659X</orcidid></addata></record>
fulltext fulltext
identifier ISSN: 1573-7721
ispartof Multimedia tools and applications, 2024-02, Vol.83 (31), p.76991-77007
issn 1573-7721
1380-7501
1573-7721
language eng
recordid cdi_proquest_journals_3100357016
source SpringerLink Journals - AutoHoldings
subjects Artificial neural networks
Computer Communication Networks
Computer Science
Constraints
Data augmentation
Data Structures and Information Theory
Deception
Deepfake
Embedding
Feature extraction
Machine learning
Multimedia Information Systems
Representations
Similarity
Special Purpose and Application-Based Systems
Track 6: Computer Vision for Multimedia Applications
title ESRL: efficient similarity representation learning for deepfake detection
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T21%3A27%3A24IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=ESRL:%20efficient%20similarity%20representation%20learning%20for%20deepfake%20detection&rft.jtitle=Multimedia%20tools%20and%20applications&rft.au=Wang,%20Feng&rft.date=2024-02-21&rft.volume=83&rft.issue=31&rft.spage=76991&rft.epage=77007&rft.pages=76991-77007&rft.issn=1573-7721&rft.eissn=1573-7721&rft_id=info:doi/10.1007/s11042-024-18447-x&rft_dat=%3Cproquest_cross%3E3100357016%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3100357016&rft_id=info:pmid/&rfr_iscdi=true