Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey on Facial Image Privacy Preservation in Cloud-Based Services

Chen Chen, Mengyuan Sun, Xueluan Gong, Yanjiao Chen, Qian Wang

TL;DR

This survey addresses the privacy risks of facial data processed in cloud-based services and classifies protection strategies into image obfuscation-based and adversarial perturbation-based approaches. It surveys representative methods, contrasts their strengths and weaknesses, and provides qualitative and quantitative evaluations across realistic metrics and datasets. Key findings include the trade-offs between protection strength and visual quality, with REM often offering a balanced solution, and the importance of extending protection to video scenarios and robustly handling real-world variations. The work highlights open challenges—such as bias, artifact reduction, and defense against augmentation or adversarial training—while outlining concrete future directions to advance practical privacy preservation in cloud environments.

Abstract

Facial recognition models are increasingly employed by commercial enterprises, government agencies, and cloud service providers for identity verification, consumer services, and surveillance. These models are often trained using vast amounts of facial data processed and stored in cloud-based platforms, raising significant privacy concerns. Users' facial images may be exploited without their consent, leading to potential data breaches and misuse. This survey presents a comprehensive review of current methods aimed at preserving facial image privacy in cloud-based services. We categorize these methods into two primary approaches: image obfuscation-based protection and adversarial perturbation-based protection. We provide an in-depth analysis of both categories, offering qualitative and quantitative comparisons of their effectiveness. Additionally, we highlight unresolved challenges and propose future research directions to improve privacy preservation in cloud computing environments.

A Survey on Facial Image Privacy Preservation in Cloud-Based Services

TL;DR

This survey addresses the privacy risks of facial data processed in cloud-based services and classifies protection strategies into image obfuscation-based and adversarial perturbation-based approaches. It surveys representative methods, contrasts their strengths and weaknesses, and provides qualitative and quantitative evaluations across realistic metrics and datasets. Key findings include the trade-offs between protection strength and visual quality, with REM often offering a balanced solution, and the importance of extending protection to video scenarios and robustly handling real-world variations. The work highlights open challenges—such as bias, artifact reduction, and defense against augmentation or adversarial training—while outlining concrete future directions to advance practical privacy preservation in cloud environments.

Abstract

Facial recognition models are increasingly employed by commercial enterprises, government agencies, and cloud service providers for identity verification, consumer services, and surveillance. These models are often trained using vast amounts of facial data processed and stored in cloud-based platforms, raising significant privacy concerns. Users' facial images may be exploited without their consent, leading to potential data breaches and misuse. This survey presents a comprehensive review of current methods aimed at preserving facial image privacy in cloud-based services. We categorize these methods into two primary approaches: image obfuscation-based protection and adversarial perturbation-based protection. We provide an in-depth analysis of both categories, offering qualitative and quantitative comparisons of their effectiveness. Additionally, we highlight unresolved challenges and propose future research directions to improve privacy preservation in cloud computing environments.
Paper Structure (18 sections, 4 figures, 2 tables)

This paper contains 18 sections, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Preliminaries
  3. Face Recognition
  4. Cloud-Based Services and Facial Data Processing
  5. State-of-the-art Face Image Protection Approaches
  6. Image Obfuscation-based Protection Methods
  7. Specific properties obfuscation
  8. Whole image obfuscation
  9. Adversarial Perturbation-based Protection Methods
  10. Non-auxiliary data requirement
  11. Auxiliary data requirement
  12. Comparison and Evaluation
  13. Qualitative Comparison
  14. Performance Evaluation
  15. Future Research Directions
  16. ...and 3 more sections

Figures (4)

  • Figure 1: Protecting private facial privacy from unauthorized exploitation in cloud-based services.
  • Figure 2: Comparison of facial privacy protection performance of EM huang2021unlearnable, REM fu2021robust, and OPOM zhong2022opom.
  • Figure 3: Comparison of facial privacy protection performance of APF zhang2020adversarial, Fawkes shan2020fawkes, and TIP-IM yang2021towards.
  • Figure 4: Visual comparison results of different facial privacy protection methods.