Abstract
Distributed multimedia surveillance systems utilize heterogeneous sensors such as cameras, motion sensors, sound sensors, and RFID in order to provide safety and security to people. However, due to the potential of exposing privacy by these systems, many people are reluctant to be electronically monitored and suffer from privacy loss. In order to overcome this dilemma, the current surveillance systems should adopt improved privacy preservation (i.e. hiding people’s face) mechanism while they are used for typical surveillance tasks. This paper takes a holistic approach to identify the different privacy leakage channels in the distributed video surveillance context and proposes the design of a secure privacy vault to conceal privacy-sensitive data obtained from distributed visual sensors. It also shows how the proposed solution helps to mitigate the potential privacy leakage problems at different levels of the leakage channels. In order to demonstrate the viability of the proposed approach, we further provide the privacy leakage attack model as well as the security analysis of the proposed solution.
•Identification of privacy leakage channels by means of privacy leakage trees.•A secure privacy vault design for distributed surveillance system.•Robust against different security and privacy attacks.•Different attack models (White-box, Gray-box, or Black-box) have been considered.