Privacy-preserving identification for monitoring images
Publication Type
Book Chapter
Publication Date
1-2024
Abstract
Camera sensors embedded in monitor units or mobile phones make it easy to capture various personal images in daily life. Machine learning especially deep learning provides an elegant way to identify images (e.g., person re-identification, face recognition, facial expression recognition). However, a personal image usually involves an amount of sensitive data, such as identity, face, and facial expression. Accordingly, image identification poses severe challenges of privacy leakage for persons' identities, face data, facial expressions, etc. Either GDPR (General Data Protection Regulation) or EDPS (European Data Protection Supervisor) stipulates that monitoring images involve private data and are easy to intrude on the fundamental right to privacy. In this chapter, we first sort out the privacy concerns in monitoring image identification and then formalize privacy-preserving identification for monitoring images. Next, we give a general framework to achieve privacy-preserving monitoring image identification and discuss privacy-preserving person re-identification based on the proposed framework. Finally, we conclude the research challenges and attempt to foresee some new research directions in privacy-preserving monitoring image identification.
Discipline
Graphics and Human Computer Interfaces | Information Security
Publication
Access Control and Security Monitoring of Multimedia Information Processing and Transmission
Editor
Zhihan Lyu, J. Lloret, & H. H. Song
ISBN
9781839536939
Identifier
10.1049/PBPC061E_ch10
Publisher
Institution of Engineering and Technology
City or Country
London
Citation
ZHAO, Bowen and LI, Xiaoguo.
Privacy-preserving identification for monitoring images. (2024). Access Control and Security Monitoring of Multimedia Information Processing and Transmission.
Available at: https://ink.library.smu.edu.sg/sis_research/8647
Additional URL
https://doi.org/10.1049/PBPC061E_ch10