Generating Master Faces for Use in Performing Wolf Attacks on Face Recognition Systems

Nguyen, Huy H.; Yamagishi, Junichi; Echizen, Isao; Marcel, Sébastien

doi:10.1109/ijcb48548.2020.9304893

Cited by 21 publications

(33 citation statements)

References 43 publications

(101 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This allows better control of the image synthesis process and results in the generation of quality and more detailed images. c) Master sample attacks: A parallel work that we became aware of post-publication [27] attempts to generates faces that are similar to a large portion of the faces in a given dataset using StyleGAN but with a different evolutionary strategy, which is one of our baselines (CMA-ES). In the field of finger print verification, Roy et al [33] have suggested to exploiting the lower quality of partial fingerprint systems to generate fingerprint templates that can be matched to a large number of users' fingerprints, without any knowledge of the actual user.…”

Section: Related Workmentioning

confidence: 99%

Generating Master Faces for Dictionary Attacks with a Network-Assisted Latent Space Evolution

Shmelkin¹,

Friedlander²,

Wolf³

2021

Preprint

View full text Add to dashboard Cite

A master face is a face image that passes facebased identity-authentication for a large portion of the population. These faces can be used to impersonate, with a high probability of success, any user, without having access to any user-information. We optimize these faces, by using an evolutionary algorithm in the latent embedding space of the StyleGAN face generator. Multiple evolutionary strategies are compared, and we propose a novel approach that employs a neural network in order to direct the search in the direction of promising samples, without adding fitness evaluations. The results we present demonstrate that it is possible to obtain a high coverage of the LFW identities (over 40%) with less than 10 master faces, for three leading deep face recognition systems.

show abstract

Section: Related Workmentioning

confidence: 99%

Generating Master Faces for Dictionary Attacks with a Network-Assisted Latent Space Evolution

Shmelkin¹,

Friedlander²,

Wolf³

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…They worked with small sensors of smartphones which acquire a lot of partial minutiae-based feature vectors. It was later improved with machine learning based approaches in [31,30,4] and recently applied on face recognition in [23].…”

Section: Definitions and Related Workmentioning

confidence: 99%

Biometric Masterkeys

Gernot,

Lacharme

2021

Preprint

View full text Add to dashboard Cite

Biometric authentication is used to secure digital or physical access. Such an authentication system uses a biometric database, where data are sometimes protected by cancelable transformations. This paper introduces the notion of biometric masterkeys. A masterkey is a feature vector such that the corresponding template matches with a significant number of templates stored in a cancelable biometric database. Such a masterkey is directly researched from a cancelable biometric database, but we also investigate another scenario in which the masterkey is fixed before the creation of the cancelable biometric database, providing additional access rights in the system for the masterkey's owner. Experimental results on the fingerprint database FVC and the face image database LFW show the effectiveness and the efficiency of such masterkeys in both scenarios. In particular, from any given feature vector, we are able to construct a cancelable database, for which the biometric template matches with all the templates of the database.

show abstract

“…In 3D, this drops to 7.5% indicating that for higher-dimensional representations the effectiveness might further decrease. ate any user without having access to information about the enrolled subject [17]. Since face recognition systems (FRS) are spreading worldwide and are increasingly involved in our daily life activities [26], successful MasterFace attacks pose a great threat to these systems.…”

Section: Introductionmentioning

confidence: 99%

“…In that work, the authors report that a single face can cover more than 20% of the identities in the used database (for a given FR model at a threshold for a false match rate of 10 −3 ). Also, Nguyen et al [17,16] proposed a similar method for generating MasterFaces and demonstrated that these attacks can successfully compromise FRS. However, these works (a) make use of older FRS, (b) consist of only limited cross-dataset and cross-model evaluations, and (c) conduct their experiments on testing data of small size and variance.…”

Section: Introductionmentioning

confidence: 99%

“…Contrarily to previous works, our experiments (a) make use of six stateof-the-art FR models, (b) are based on cross-dataset and cross-model evaluations, and (c) utilize testing datasets that are of a significantly higher size and variance. To ensure comparability with previous works [23,17], we use the original MasterFace images reported in the Arxiv versions of these papers (CC-BY-4.0) for the experiments. The empirical investigations showed that only an insignificant proportion of the testing identities are successfully matched against the MasterFaces and that the comparison scores resulting from comparisons with MasterFaces closely resembles the score distribution of zero-effort imposter.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the (Limited) Generalization of MasterFace Attacks and Its Relation to the Capacity of Face Representations

Terhörst¹,

Bierbaum²,

Huber³

et al. 2022

Preprint

View full text Add to dashboard Cite

A MasterFace is a face image that can successfully match against a large portion of the population. Since their generation does not require access to the information of the enrolled subjects, MasterFace attacks represent a potential security risk for widely-used face recognition systems. Previous works proposed methods for generating such images and demonstrated that these attacks can strongly compromise face recognition. However, previous works followed evaluation settings consisting of older recognition models, limited cross-dataset and cross-model evaluations, and the use of low-scale testing data. This makes it hard to state the generalizability of these attacks. In this work, we comprehensively analyse the generalizability of MasterFace attacks in empirical and theoretical investigations. The empirical investigations include the use of six state-of-the-art FR models, cross-dataset and cross-model evaluation protocols, and utilizing testing datasets of significantly higher size and variance. The results indicate a low generalizability when MasterFaces are training on a different face recognition model than the one used for testing. In these cases, the attack performance is similar to zero-effort imposter attacks. In the theoretical investigations, we define and estimate the face capacity and the maximum MasterFace coverage under the assumption that identities in the face space are well separated. The current trend of increasing the fairness and generalizability in face recognition indicates that the vulnerability of future systems might further decrease. Future works might analyse the utility of MasterFaces for understanding and enhancing the robustness of face recognition models.

show abstract

Generating Master Faces for Use in Performing Wolf Attacks on Face Recognition Systems

Cited by 21 publications

References 43 publications

Generating Master Faces for Dictionary Attacks with a Network-Assisted Latent Space Evolution

Generating Master Faces for Dictionary Attacks with a Network-Assisted Latent Space Evolution

Biometric Masterkeys

On the (Limited) Generalization of MasterFace Attacks and Its Relation to the Capacity of Face Representations

Contact Info

Product

Resources

About