The super-recogniser advantage extends to the detection of digitally manipulated faces

Author:

Davis Josh PORCID,Robertson David JORCID,Jenkins Ryan,Nichols Robert,Ibsen MathiasORCID,Babbs Martha,Rathgeb Christian,Løvåsdal Frøy,Raja Kiran Bylappa,Busch Christoph

Abstract

Face recognition by human officials remains the predominant method of identity verification in security-critical contexts (e.g., passport renewal, border control). The integrity of this process can be compromised by sophisticated fraud attacks using digitally manipulated face images. Therefore, in this study we examine whether human observers can robustly detect digitally manipulated passport photos and whether super-recognisers (SRs), individuals who excel at identity recognition, outperform typical recogniser controls. Using two face manipulation detection tasks (DFMD1, DFMD2), participants were asked to decide whether a ‘suspected’ passport photo had been digitally manipulated. The findings show that while both groups could consistently detect these manipulated faces, SRs significantly outperformed controls. This effect was not the result of a ‘speed-accuracy’ trade off, and face identification task performance and self-rated face recognition aptitude significantly predicted 30% of the variance in manipulated image detection scores. Taken together, these findings show that, despite increasing sophistication in digital face manipulation techniques, there is still utility in employing human operators, particularly SRs, to detect them. Further cognitive and applied implications of these findings are discussed.

Publisher

Center for Open Science

Cited by 1 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Conditional Face Image Manipulation Detection: Combining Algorithm and Human Examiner Decisions;Proceedings of the 2024 ACM Workshop on Information Hiding and Multimedia Security;2024-06-24

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3