Privacy-preserving Secure Media Streaming for Multi-user Smart Environments

Over the last years, our lifestyle has been positively upset by the sudden advent of technology. The Internet of Things (IoT), offering universal and ubiquitous connectivity to both people and objects, revealed to be the silver bullet for enabling a vast number of previously unexpected applications. In particular, media streaming providers are growing in business and scope, and we can forecast that soon, video streaming will substitute TV broadcasting activities. With the increasing success of multi-user smart environments, empowered by new-generation smart devices and IoT architectures, multimedia contents (i.e., images and videos) need to be effectively accessed anytime and anywhere. Recent advances in computer vision technologies have made the development of intelligent monitoring systems for video surveillance and ambient-assisted living. Such a scenario permits better integration among technologies, multimedia content, and end-users. However, there are several challenges, and some are still open. More precisely, due to the sensitivity of some multimedia content (e.g., video-surveillance streams), it is paramount to preserve users' privacy. Again, it is necessary to guarantee the integrity of usage rights during any multimedia transmission process, starting from the video encoding phase. In this way, the private content is disclosed only when the stream is decoded on the other endpoint, by the legitimate user. In this article, we present a secure video transmission strategy that can address the challenges mentioned above. The proposed strategy takes advantage of both watermarking and video scrambling techniques to make it possible for the secure and privacy-preserving transmission of multimedia streaming. Through our proposal, multimedia streaming is of low quality and thus unusable. However, it can be fully recovered and enjoyed only by authorized users. Finally, due to its low complexity and energy-efficiency, our proposal is particularly suitable for onboard implementations.