Face Recognition Accuracy of Forensic Examiners, Superrecognizers, and Face Recognition Algorithms

Simple item page
dc.contributor.authorPhillips, P. J.
dc.contributor.authorYates, A. N.
dc.contributor.authorHu, Ying
dc.contributor.authorHahn, Carina A.
dc.contributor.authorNoyes, Eilidh
dc.contributor.authorJackson, Kelsey
dc.contributor.authorCavazos, Jacqueline G.
dc.contributor.authorJeckeln, Géraldine
dc.contributor.authorRanjan, R.
dc.contributor.authorSankaranarayanan, S.
dc.contributor.authorChen, J. -C
dc.contributor.authorCastillo, C. D.
dc.contributor.authorChellappa, R.
dc.contributor.authorWhite, D.
dc.contributor.authorO'Toole, Alice J.
dc.contributor.utdAuthorHu, Ying
dc.contributor.utdAuthorHahn, Carina A.
dc.contributor.utdAuthorNoyes, Eilidh
dc.contributor.utdAuthorJackson, Kelsey
dc.contributor.utdAuthorCavazos, Jacqueline G.
dc.contributor.utdAuthorJeckeln, Géraldine
dc.contributor.utdAuthorO'Toole, Alice J.
dc.date.accessioned2019-05-31T22:54:22Z
dc.date.available2019-05-31T22:54:22Z
dc.date.created2018
dc.descriptionIncludes supplementary material
dc.description.abstractAchieving the upper limits of face identification accuracy in forensic applications can minimize errors that have profound social and personal consequences. Although forensic examiners identify faces in these applications, systematic tests of their accuracy are rare. How can we achieve the most accurate face identification: using people and/or machines working alone or in collaboration? In a comprehensive comparison of face identification by humans and computers, we found that forensic facial examiners, facial reviewers, and superrecognizers were more accurate than fingerprint examiners and students on a challenging face identification test. Individual performance on the test varied widely. On the same test, four deep convolutional neural networks (DCNNs), developed between 2015 and 2017, identified faces within the range of human accuracy. Accuracy of the algorithms increased steadily over time, with the most recent DCNN scoring above the median of the forensic facial examiners. Using crowd-sourcing methods, we fused the judgments of multiple forensic facial examiners by averaging their rating-based identity judgments. Accuracy was substantially better for fused judgments than for individuals working alone. Fusion also served to stabilize performance, boosting the scores of lower-performing individuals and decreasing variability. Single forensic facial examiners fused with the best algorithm were more accurate than the combination of two examiners. Therefore, collaboration among humans and between humans and machines offers tangible benefits to face identification accuracy in important applications. These results offer an evidence-based roadmap for achieving the most accurate face identification possible.
dc.description.departmentSchool of Behavioral and Brain Sciences
dc.description.sponsorshipIntelligence Advanced Research Projects Activity (IARPA) via IARPA R&D Contract 2014-14071600012; Australian Research Council Linkage Projects LP160101523 and LP130100702; and National Institute of Justice Grant 2015-IJ-CX-K014.
dc.identifier.bibliographicCitationPhillips, P. J., A. N. Yates, Y. Hu, C. A. Hahn, et al. 2018. "Face recognition accuracy of forensic examiners, superrecognizers, and face recognition algorithms." Proceedings of the National Academy of Sciences of the United States of America 115(24): 6171-6176, doi:10.1073/pnas.1721355115
dc.identifier.issn0027-8424
dc.identifier.issue24
dc.identifier.urihttps://hdl.handle.net/10735.1/6556
dc.identifier.volume115
dc.language.isoen
dc.publisherNational Academy of Sciences
dc.relation.urihttp://dx.doi.org/10.1073/pnas.1721355115
dc.rightsCC BY-NC-ND 4.0 (Attribution-NonCommercial-NoDerivatives)
dc.rights©2018 The Authors
dc.rights.urihttps://creativecommons.org/licenses/by-nc-nd/4.0/
dc.source.journalProceedings of the National Academy of Sciences of the United States of America
dc.subjectFace recognition, Human (Computer science)
dc.subjectForensic sciences
dc.subjectMachine learning
dc.subjectOptical pattern recognition
dc.subjectIdentification
dc.subjectPersons
dc.titleFace Recognition Accuracy of Forensic Examiners, Superrecognizers, and Face Recognition Algorithms
dc.type.genrearticle

Files

Original bundle

Now showing 1 - 3 of 3
Thumbnail Image
Name:
BBS-6436-9443.36.pdf
Size:
1.16 MB
Format:
Adobe Portable Document Format
Description:
Article
Download
Thumbnail Image
Name:
BBS-6436-9443.36_S1.pdf
Size:
4.65 MB
Format:
Adobe Portable Document Format
Description:
Supplement
Download
No Thumbnail Available
Name:
BBS-6436-9443.36_S2.zip
Size:
4.88 KB
Format:
Unknown data format
Description:
Data Sets
Download

Collections

O'Toole, Alice J.