A Momentum-Based Local Face Adversarial Example Generation Algorithm
Small perturbations can make deep models fail. Since deep models are widely used in face recognition systems (FRS) such as surveillance and access control, adversarial examples may introduce more subtle threats to face recognition systems. In this paper, we propose a practical white-box adversarial...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-12-01
|
Series: | Algorithms |
Subjects: | |
Online Access: | https://www.mdpi.com/1999-4893/15/12/465 |
_version_ | 1797461832634466304 |
---|---|
author | Dapeng Lang Deyun Chen Jinjie Huang Sizhao Li |
author_facet | Dapeng Lang Deyun Chen Jinjie Huang Sizhao Li |
author_sort | Dapeng Lang |
collection | DOAJ |
description | Small perturbations can make deep models fail. Since deep models are widely used in face recognition systems (FRS) such as surveillance and access control, adversarial examples may introduce more subtle threats to face recognition systems. In this paper, we propose a practical white-box adversarial attack method. The method can automatically form a local area with key semantics on the face. The shape of the local area generated by the algorithm varies according to the environment and light of the character. Since these regions contain major facial features, we generated patch-like adversarial examples based on this region, which can effectively deceive FRS. The algorithm introduced the momentum parameter to stabilize the optimization directions. We accelerated the generation process by increasing the learning rate in segments. Compared with the traditional adversarial algorithm, our algorithms are very inconspicuous, which is very suitable for application in real scenes. The attack was verified on the CASIA WebFace and LFW datasets which were also proved to have good transferability. |
first_indexed | 2024-03-09T17:24:56Z |
format | Article |
id | doaj.art-6c3d0970d9454a8987d8c2a3d03a6a9a |
institution | Directory Open Access Journal |
issn | 1999-4893 |
language | English |
last_indexed | 2024-03-09T17:24:56Z |
publishDate | 2022-12-01 |
publisher | MDPI AG |
record_format | Article |
series | Algorithms |
spelling | doaj.art-6c3d0970d9454a8987d8c2a3d03a6a9a2023-11-24T12:49:21ZengMDPI AGAlgorithms1999-48932022-12-01151246510.3390/a15120465A Momentum-Based Local Face Adversarial Example Generation AlgorithmDapeng Lang0Deyun Chen1Jinjie Huang2Sizhao Li3School of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080, ChinaSchool of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080, ChinaSchool of Computer Science and Technology, Harbin University of Science and Technology, Harbin 150080, ChinaCollege of Computer Science and Technology, Harbin Engineering University, Harbin 150001, ChinaSmall perturbations can make deep models fail. Since deep models are widely used in face recognition systems (FRS) such as surveillance and access control, adversarial examples may introduce more subtle threats to face recognition systems. In this paper, we propose a practical white-box adversarial attack method. The method can automatically form a local area with key semantics on the face. The shape of the local area generated by the algorithm varies according to the environment and light of the character. Since these regions contain major facial features, we generated patch-like adversarial examples based on this region, which can effectively deceive FRS. The algorithm introduced the momentum parameter to stabilize the optimization directions. We accelerated the generation process by increasing the learning rate in segments. Compared with the traditional adversarial algorithm, our algorithms are very inconspicuous, which is very suitable for application in real scenes. The attack was verified on the CASIA WebFace and LFW datasets which were also proved to have good transferability.https://www.mdpi.com/1999-4893/15/12/465adversarial examplesface recognitionmask matrixtargeted attacknon-targeted attack |
spellingShingle | Dapeng Lang Deyun Chen Jinjie Huang Sizhao Li A Momentum-Based Local Face Adversarial Example Generation Algorithm Algorithms adversarial examples face recognition mask matrix targeted attack non-targeted attack |
title | A Momentum-Based Local Face Adversarial Example Generation Algorithm |
title_full | A Momentum-Based Local Face Adversarial Example Generation Algorithm |
title_fullStr | A Momentum-Based Local Face Adversarial Example Generation Algorithm |
title_full_unstemmed | A Momentum-Based Local Face Adversarial Example Generation Algorithm |
title_short | A Momentum-Based Local Face Adversarial Example Generation Algorithm |
title_sort | momentum based local face adversarial example generation algorithm |
topic | adversarial examples face recognition mask matrix targeted attack non-targeted attack |
url | https://www.mdpi.com/1999-4893/15/12/465 |
work_keys_str_mv | AT dapenglang amomentumbasedlocalfaceadversarialexamplegenerationalgorithm AT deyunchen amomentumbasedlocalfaceadversarialexamplegenerationalgorithm AT jinjiehuang amomentumbasedlocalfaceadversarialexamplegenerationalgorithm AT sizhaoli amomentumbasedlocalfaceadversarialexamplegenerationalgorithm AT dapenglang momentumbasedlocalfaceadversarialexamplegenerationalgorithm AT deyunchen momentumbasedlocalfaceadversarialexamplegenerationalgorithm AT jinjiehuang momentumbasedlocalfaceadversarialexamplegenerationalgorithm AT sizhaoli momentumbasedlocalfaceadversarialexamplegenerationalgorithm |