Face Database Protection via Beautification with Chaotic Systems
The database of faces containing sensitive information is at risk of being targeted by unauthorized automatic recognition systems, which is a significant concern for privacy. Although there are existing methods that aim to conceal identifiable information by adding adversarial perturbations to faces...
| Huvudupphovsmän: | , , |
|---|---|
| Materialtyp: | Artikel |
| Språk: | English |
| Publicerad: |
MDPI AG
2023-03-01
|
| Serie: | Entropy |
| Ämnen: | |
| Länkar: | https://www.mdpi.com/1099-4300/25/4/566 |
| _version_ | 1827745193345941504 |
|---|---|
| author | Tao Wang Yushu Zhang Ruoyu Zhao |
| author_facet | Tao Wang Yushu Zhang Ruoyu Zhao |
| author_sort | Tao Wang |
| collection | DOAJ |
| description | The database of faces containing sensitive information is at risk of being targeted by unauthorized automatic recognition systems, which is a significant concern for privacy. Although there are existing methods that aim to conceal identifiable information by adding adversarial perturbations to faces, they suffer from noticeable distortions that significantly compromise visual perception, and therefore, offer limited protection to privacy. Furthermore, the increasing prevalence of appearance anxiety on social media has led to users preferring to beautify their faces before uploading images. In this paper, we design a novel face database protection scheme via beautification with chaotic systems. Specifically, we construct the adversarial face with better visual perception via beautification for each face in the database. In the training, the face matcher and the beautification discriminator are federated against the generator, prompting it to generate beauty-like perturbations on the face to confuse the face matcher. Namely, the pixel changes produced by face beautification mask the adversarial perturbations. Moreover, we use chaotic systems to disrupt the order of adversarial faces in the database, further mitigating the risk of privacy leakage. Our scheme has been extensively evaluated through experiments, which show that it effectively defends against unauthorized attacks while also yielding good visual results. |
| first_indexed | 2024-03-11T05:03:38Z |
| format | Article |
| id | doaj.art-f4af2824dc3a45159a41313a8be2c425 |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-03-11T05:03:38Z |
| publishDate | 2023-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-f4af2824dc3a45159a41313a8be2c4252023-11-17T19:07:53ZengMDPI AGEntropy1099-43002023-03-0125456610.3390/e25040566Face Database Protection via Beautification with Chaotic SystemsTao Wang0Yushu Zhang1Ruoyu Zhao2College of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, ChinaCollege of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, ChinaCollege of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, ChinaThe database of faces containing sensitive information is at risk of being targeted by unauthorized automatic recognition systems, which is a significant concern for privacy. Although there are existing methods that aim to conceal identifiable information by adding adversarial perturbations to faces, they suffer from noticeable distortions that significantly compromise visual perception, and therefore, offer limited protection to privacy. Furthermore, the increasing prevalence of appearance anxiety on social media has led to users preferring to beautify their faces before uploading images. In this paper, we design a novel face database protection scheme via beautification with chaotic systems. Specifically, we construct the adversarial face with better visual perception via beautification for each face in the database. In the training, the face matcher and the beautification discriminator are federated against the generator, prompting it to generate beauty-like perturbations on the face to confuse the face matcher. Namely, the pixel changes produced by face beautification mask the adversarial perturbations. Moreover, we use chaotic systems to disrupt the order of adversarial faces in the database, further mitigating the risk of privacy leakage. Our scheme has been extensively evaluated through experiments, which show that it effectively defends against unauthorized attacks while also yielding good visual results.https://www.mdpi.com/1099-4300/25/4/566database protectionadversarialbeautificationchaotic systems |
| spellingShingle | Tao Wang Yushu Zhang Ruoyu Zhao Face Database Protection via Beautification with Chaotic Systems Entropy database protection adversarial beautification chaotic systems |
| title | Face Database Protection via Beautification with Chaotic Systems |
| title_full | Face Database Protection via Beautification with Chaotic Systems |
| title_fullStr | Face Database Protection via Beautification with Chaotic Systems |
| title_full_unstemmed | Face Database Protection via Beautification with Chaotic Systems |
| title_short | Face Database Protection via Beautification with Chaotic Systems |
| title_sort | face database protection via beautification with chaotic systems |
| topic | database protection adversarial beautification chaotic systems |
| url | https://www.mdpi.com/1099-4300/25/4/566 |
| work_keys_str_mv | AT taowang facedatabaseprotectionviabeautificationwithchaoticsystems AT yushuzhang facedatabaseprotectionviabeautificationwithchaoticsystems AT ruoyuzhao facedatabaseprotectionviabeautificationwithchaoticsystems |