Papers › A2-LINK: Recognizing Disguised Faces via Active Learning and Adversarial Noise based...

A2-LINK: Recognizing Disguised Faces via Active Learning and Adversarial Noise based Inter-Domain Knowledge

1 Jun 2020IEEE Transactions on Biometrics, Behavior, and Identity Science 2020 6archive 2025-07-28

Anshuman Suri, Mayank Vatsa, Richa Singh

Face recognition in the unconstrained environment is an ongoing research challenge. Although several covariates of face recognition such as pose and low resolution have received significant attention“, disguise” is considered an onerous covariate of face recognition. One of the primary reasons for this is the scarcity of large and representative labeled databases, along with the lack of algorithms that work well for multiple covariates in such environments. In order to address the problem of face recognition in the presence of disguise, the paper proposes an active learning framework termed as A2-LINK. Starting with a face recognition machine-learning model, A2-LINK intelligently selects training samples from the target domain to be labeled and, using hybrid noises such as adversarial noise, fine-tunes a model that works well both in the presence and absence of disguise. Experimental results demonstrate the effectiveness and generalization of the proposed framework on the DFW and DFW2019 datasets with state-of-the-art deep learning featurization models such as LCSSE, ArcFace, and DenseNet.

PaperPDFCode

Code

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Active LearningDomain AdaptationFace RecognitionHeterogeneous Face Recognition

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Heterogeneous Face Recognition Disguised Faces in the Wild A2-LINK GAR @0.1% FAR Impersonation 69.27 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild A2-LINK GAR @0.1% FAR Obfuscation 93.08 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild A2-LINK GAR @0.1% FAR Overall 93.01 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild A2-LINK GAR @1% FAR Impersonation 99.01 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild A2-LINK GAR @1% FAR Obfuscation 95.93 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild A2-LINK GAR @1% FAR Overall 95.99 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.01% FAR Impersonation 54.40 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.01% FAR Obfuscation 97.20 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.01% FAR Overall 96.18 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.01% FAR Plastic Surgery 96.00 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.1% FAR Impersonation 79.2 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.1% FAR Obfuscation 99.00 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.1% FAR Overall 98.63 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A2-LINK GAR @0.1% FAR Plastic Surgery 98.80 #1 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.01% FAR Impersonation 52.80 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.01% FAR Obfuscation 94.02 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.01% FAR Overall 93.06 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.01% FAR Plastic Surgery 92.00 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.1% FAR Impersonation 76.40 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.1% FAR Obfuscation 96.84 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.1% FAR Overall 95.96 #2 of 2 Archive leaderboard report
Heterogeneous Face Recognition Disguised Faces in the Wild 2019 A-LINK GAR @0.1% FAR Plastic Surgery 95.20 #2 of 2 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 ConvolutionArcFaceAverage PoolingBatch NormalizationConcatenated Skip ConnectionConvolutionDense BlockDense ConnectionsDropoutGlobal Average PoolingKaiming InitializationMax PoolingReLUSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections