Sensor-level security and privacy protection by embedding video content analysis

In traditional video camera networks privacy protection is performed, if at all, at the data center after the video has been streamed over the network. With the on-board processing capabilities of today´s visual sensor network devices, image processing and privacy protection can be performed now on the cameras before any data is sent over the network. While moving privacy protection closer to the data source is a big step forward, it raises also new challenges. The growing software stack of today´s embedded devices and the widespread use of public wired and wireless network infrastructure entail substantial effort to secure this new class of devices. In this position paper we describe a novel approach that aims to tightly integrate security and privacy protection with the image sensor itself and thereby eliminating many security issues of today´s embedded camera systems. Resource-efficient, online video content analysis plays a vital role in this endeavor towards making privacy protection an inherent property of future image acquisition units. This paper outlines the requirements and challenges of such an approach and discusses potential implementation strategies.