Lightweight and Practical Privacy-Preserving Image Masking in Smart Community

Abstract

Advances in the Internet of Things (IoT) and telecommunications technologies (e.g., 5G) have contributed to the development of smart cities and nations (collectively referred to as smart communities). In a smart community, IoT devices can collect significant information about urban residents (e.g., a large number of images collected by cameras containing sensitive information), and such information may be shared with intermediate nodes. In real-world deployment, intermediate nodes are not completely trusted, where the information collected may be used for commercial purposes (e.g., user profiling and advertising) or malicious activities (e.g., covert surveillance). In this paper, we introduce an approach to ensure privacy-preserving image masking. Specifically, before the image is transmitted to the camera owner or the monitoring cloud platform, only sensitive areas instead of the entire image will be processed according to the camera owner’s settings, which allows to significantly reduce the computational cost. Then, in order to reduce the interactions between the community data center and the IoT camera, the monitoring cloud platform performs proxy re-encryption. This allows the community data center to recover the original image without relying on the IoT device’s private key. Our evaluation indicates the utility and efficiency of our approach, as compared with similar schemes.