Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach

In recent years, the multimedia forensics and security community has seen remarkable progress in multitask learning for DeepFake (i.e., face forgery) detection. The prevailing strategy has been to frame DeepFake detection as a binary classification problem augmented by manipulation-oriented auxiliar...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zou, Mian, Yu, Baosheng, Zhan, Yibing, Lyu, Siwei, Ma, Kede
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Zou, Mian
Yu, Baosheng
Zhan, Yibing
Lyu, Siwei
Ma, Kede
description In recent years, the multimedia forensics and security community has seen remarkable progress in multitask learning for DeepFake (i.e., face forgery) detection. The prevailing strategy has been to frame DeepFake detection as a binary classification problem augmented by manipulation-oriented auxiliary tasks. This strategy focuses on learning features specific to face manipulations, which exhibit limited generalizability. In this paper, we delve deeper into semantics-oriented multitask learning for DeepFake detection, leveraging the relationships among face semantics via joint embedding. We first propose an automatic dataset expansion technique that broadens current face forgery datasets to support semantics-oriented DeepFake detection tasks at both the global face attribute and local face region levels. Furthermore, we resort to joint embedding of face images and their corresponding labels (depicted by textual descriptions) for prediction. This approach eliminates the need for manually setting task-agnostic and task-specific parameters typically required when predicting labels directly from images. In addition, we employ a bi-level optimization strategy to dynamically balance the fidelity loss weightings of various tasks, making the training process fully automated. Extensive experiments on six DeepFake datasets show that our method improves the generalizability of DeepFake detection and, meanwhile, renders some degree of model interpretation by providing human-understandable explanations.
doi_str_mv 10.48550/arxiv.2408.16305
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2408_16305</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2408_16305</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2408_163053</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjGw0DM0MzYw5WSICE7NTcwryUwu1vUvykzNK0lNUfAtzSnJLEkszlbwSU0sysvMS1dIyy9ScElNLXBLzE4FMkpSk0sy8_OsFBwVvPIz80oUXHOTUlNSQCodCwqK8hOTM3gYWNMSc4pTeaE0N4O8m2uIs4cu2A3xBUWZuYlFlfEgt8SD3WJMWAUA9i8_Bg</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach</title><source>arXiv.org</source><creator>Zou, Mian ; Yu, Baosheng ; Zhan, Yibing ; Lyu, Siwei ; Ma, Kede</creator><creatorcontrib>Zou, Mian ; Yu, Baosheng ; Zhan, Yibing ; Lyu, Siwei ; Ma, Kede</creatorcontrib><description>In recent years, the multimedia forensics and security community has seen remarkable progress in multitask learning for DeepFake (i.e., face forgery) detection. The prevailing strategy has been to frame DeepFake detection as a binary classification problem augmented by manipulation-oriented auxiliary tasks. This strategy focuses on learning features specific to face manipulations, which exhibit limited generalizability. In this paper, we delve deeper into semantics-oriented multitask learning for DeepFake detection, leveraging the relationships among face semantics via joint embedding. We first propose an automatic dataset expansion technique that broadens current face forgery datasets to support semantics-oriented DeepFake detection tasks at both the global face attribute and local face region levels. Furthermore, we resort to joint embedding of face images and their corresponding labels (depicted by textual descriptions) for prediction. This approach eliminates the need for manually setting task-agnostic and task-specific parameters typically required when predicting labels directly from images. In addition, we employ a bi-level optimization strategy to dynamically balance the fidelity loss weightings of various tasks, making the training process fully automated. Extensive experiments on six DeepFake datasets show that our method improves the generalizability of DeepFake detection and, meanwhile, renders some degree of model interpretation by providing human-understandable explanations.</description><identifier>DOI: 10.48550/arxiv.2408.16305</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition</subject><creationdate>2024-08</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2408.16305$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2408.16305$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Zou, Mian</creatorcontrib><creatorcontrib>Yu, Baosheng</creatorcontrib><creatorcontrib>Zhan, Yibing</creatorcontrib><creatorcontrib>Lyu, Siwei</creatorcontrib><creatorcontrib>Ma, Kede</creatorcontrib><title>Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach</title><description>In recent years, the multimedia forensics and security community has seen remarkable progress in multitask learning for DeepFake (i.e., face forgery) detection. The prevailing strategy has been to frame DeepFake detection as a binary classification problem augmented by manipulation-oriented auxiliary tasks. This strategy focuses on learning features specific to face manipulations, which exhibit limited generalizability. In this paper, we delve deeper into semantics-oriented multitask learning for DeepFake detection, leveraging the relationships among face semantics via joint embedding. We first propose an automatic dataset expansion technique that broadens current face forgery datasets to support semantics-oriented DeepFake detection tasks at both the global face attribute and local face region levels. Furthermore, we resort to joint embedding of face images and their corresponding labels (depicted by textual descriptions) for prediction. This approach eliminates the need for manually setting task-agnostic and task-specific parameters typically required when predicting labels directly from images. In addition, we employ a bi-level optimization strategy to dynamically balance the fidelity loss weightings of various tasks, making the training process fully automated. Extensive experiments on six DeepFake datasets show that our method improves the generalizability of DeepFake detection and, meanwhile, renders some degree of model interpretation by providing human-understandable explanations.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjGw0DM0MzYw5WSICE7NTcwryUwu1vUvykzNK0lNUfAtzSnJLEkszlbwSU0sysvMS1dIyy9ScElNLXBLzE4FMkpSk0sy8_OsFBwVvPIz80oUXHOTUlNSQCodCwqK8hOTM3gYWNMSc4pTeaE0N4O8m2uIs4cu2A3xBUWZuYlFlfEgt8SD3WJMWAUA9i8_Bg</recordid><startdate>20240829</startdate><enddate>20240829</enddate><creator>Zou, Mian</creator><creator>Yu, Baosheng</creator><creator>Zhan, Yibing</creator><creator>Lyu, Siwei</creator><creator>Ma, Kede</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240829</creationdate><title>Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach</title><author>Zou, Mian ; Yu, Baosheng ; Zhan, Yibing ; Lyu, Siwei ; Ma, Kede</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2408_163053</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><toplevel>online_resources</toplevel><creatorcontrib>Zou, Mian</creatorcontrib><creatorcontrib>Yu, Baosheng</creatorcontrib><creatorcontrib>Zhan, Yibing</creatorcontrib><creatorcontrib>Lyu, Siwei</creatorcontrib><creatorcontrib>Ma, Kede</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Zou, Mian</au><au>Yu, Baosheng</au><au>Zhan, Yibing</au><au>Lyu, Siwei</au><au>Ma, Kede</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach</atitle><date>2024-08-29</date><risdate>2024</risdate><abstract>In recent years, the multimedia forensics and security community has seen remarkable progress in multitask learning for DeepFake (i.e., face forgery) detection. The prevailing strategy has been to frame DeepFake detection as a binary classification problem augmented by manipulation-oriented auxiliary tasks. This strategy focuses on learning features specific to face manipulations, which exhibit limited generalizability. In this paper, we delve deeper into semantics-oriented multitask learning for DeepFake detection, leveraging the relationships among face semantics via joint embedding. We first propose an automatic dataset expansion technique that broadens current face forgery datasets to support semantics-oriented DeepFake detection tasks at both the global face attribute and local face region levels. Furthermore, we resort to joint embedding of face images and their corresponding labels (depicted by textual descriptions) for prediction. This approach eliminates the need for manually setting task-agnostic and task-specific parameters typically required when predicting labels directly from images. In addition, we employ a bi-level optimization strategy to dynamically balance the fidelity loss weightings of various tasks, making the training process fully automated. Extensive experiments on six DeepFake datasets show that our method improves the generalizability of DeepFake detection and, meanwhile, renders some degree of model interpretation by providing human-understandable explanations.</abstract><doi>10.48550/arxiv.2408.16305</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2408.16305
ispartof
issn
language eng
recordid cdi_arxiv_primary_2408_16305
source arXiv.org
subjects Computer Science - Computer Vision and Pattern Recognition
title Semantics-Oriented Multitask Learning for DeepFake Detection: A Joint Embedding Approach
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-05T06%3A35%3A49IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Semantics-Oriented%20Multitask%20Learning%20for%20DeepFake%20Detection:%20A%20Joint%20Embedding%20Approach&rft.au=Zou,%20Mian&rft.date=2024-08-29&rft_id=info:doi/10.48550/arxiv.2408.16305&rft_dat=%3Carxiv_GOX%3E2408_16305%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true