The Good, the Bad, and the Ugly Face Challenge Problem

Phillips, P. Jonathon; Beveridge, J. Ross; Draper, Bruce A.; Givens, Geof H.; O'Toole, Alice J.; Bolme, David S.; Dunlop, Joseph; Lui, Yui Man; Sahibzada, Hassan; Weimer, Samuel

doi:10.1016/j.imavis.2012.01.004

Cited by 56 publications

(49 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast, LFW-like protocol facilitates easy and fast comparison between algorithms with 6K pairs of images. Further, inspired by the ugly subset of GBU database [48], we have selected the "difficult" pairs (in term of cosine similarity) to avoid the saturated performance to be easily reported 2 .…”

Section: Racial Faces In-the-wild: Rfwmentioning

confidence: 99%

See 1 more Smart Citation

Racial Faces in the Wild: Reducing Racial Bias by Information Maximization Adaptation Network

Wang

Deng

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

291

278

View full text Add to dashboard Cite

Racial bias is an important issue in biometric, but has not been thoroughly studied in deep face recognition. In this paper, we first contribute a dedicated dataset called Racial Faces in-the-Wild (RFW) database, on which we firmly validated the racial bias of four commercial APIs and four state-of-the-art (SOTA) algorithms. Then, we further present the solution using deep unsupervised domain adaptation and propose a deep information maximization adaptation network (IMAN) to alleviate this bias by using Caucasian as source domain and other races as target domains. This unsupervised method simultaneously aligns global distribution to decrease race gap at domain-level, and learns the discriminative target representations at cluster level. A novel mutual information loss is proposed to further enhance the discriminative ability of network output without label information. Extensive experiments on RFW, GBU, and IJB-A databases show that IMAN successfully learns features that generalize well across different races and across different databases.

show abstract

Section: Racial Faces In-the-wild: Rfwmentioning

confidence: 99%

“…Ablation study shows that MI loss has unique effect on reducing racial bias. In addition, IMAN is also helpful in adapting general deep model to a specific database, and achieved improved performance on GBU [48] and IJB-A [37] databases. The contributions of this work are three aspects.…”

Section: Introductionmentioning

confidence: 99%

Racial Faces in the Wild: Reducing Racial Bias by Information Maximization Adaptation Network

Wang

Deng

et al. 2019

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

291

278

View full text Add to dashboard Cite

show abstract

“…4 prone in unconstrained environments (Phillips et al 2012). As a result, human operators are often required to compare candidate images that are suggested by the computer system.…”

Section: Crowd Effects In Unfamiliar Face Matchingmentioning

confidence: 99%

“…One possible response to this fact is to replace human viewers with automatic face recognition systems, as has begun to happen in some applied settings, notably border control. However, although there have been significant improvements in the accuracy of automatic face recognition in recent years (O'Toole, Phillips et al, 2011), it is also clear that these technologies do not work perfectly and are especially error-prone in unconstrained environments (Phillips et al, 2012). As a result, human operators are often required to compare candidate images that are suggested by the computer system.…”

Section: Introductionmentioning

confidence: 99%

Crowd Effects in Unfamiliar Face Matching

White

Burton

Kemp

et al. 2013

Applied Cognitive Psychology

View full text Add to dashboard Cite

Summary Psychological research shows that humans can not reliably match unfamiliar faces. This presents a practical problem, because identity verification processes in a variety of occupational settings depend on people to perform these tasks reliably. In this context, it is surprising that very few studies have attempted to improve human performance. Here, we investigate whether distributing face matching tasks across groups of individuals might help to solve this problem. Across four studies, we measure the accuracy of the ‘crowd’ on a standard test of face matching ability and find that aggregating individual responses produces substantial gains in matching accuracy. We discuss the practical implications of this result and also suggest ways in which this approach might be used to improve our understanding of face perception more generally. Copyright © 2013 John Wiley & Sons, Ltd.

show abstract

“…This step emphasizes the dimensions along which images of different people are spread The public domain LRPCA algorithm was introduced as part of a recently released public face recognition challenge problem. 3 While not as capable as the best engineered commercial algorithms, the LRPCA represents a significant advancement over initial PCA introduced early in this article.…”

Section: Local Region Pcamentioning

confidence: 98%

Biometric face recognition: from classical statistics to future challenges

Givens

Beveridge

Lui

et al. 2013

WIREs Computational Stats

Self Cite

View full text Add to dashboard Cite

Face recognition involves at least three major concepts from statistics: dimension reduction, feature extraction, and prediction. A selective review of algorithms, from seminal to state-of-the-art, explores how these concepts persist as organizing principles in the field. Algorithms based directly upon classical statistical techniques include linear methods like principal component analysis and linear discriminant analysis. Nonlinear manifold methods, such as Laplacianfaces and Stiefel quotients, offer considerable performance improvements. Other noteworthy ideas include three-dimensional morphable models, methods using local regions and/or alternative feature spaces (e.g., elastic bunch graph matching and local binary patterns) and sparse representation approaches. Opportunities for innovative statistical and collaborative research in face recognition are expanding in tandem with the growing complexity and diversity of applications.

show abstract

The Good, the Bad, and the Ugly Face Challenge Problem

Cited by 56 publications

References 7 publications

Racial Faces in the Wild: Reducing Racial Bias by Information Maximization Adaptation Network

Racial Faces in the Wild: Reducing Racial Bias by Information Maximization Adaptation Network

Crowd Effects in Unfamiliar Face Matching

Biometric face recognition: from classical statistics to future challenges

Contact Info

Product

Resources

About