Privacy enforcement by identity management in smart surveillance systems

Surveillance Systems have become increasingly powerful. Conventional camera-based systems are extended with different types of (wireless) sensors, the number of data sources increased and hardware as well as algorithms improved. Such smart systems do not only threaten privacy, but technological improvements can also increase privacy of observed objects. Previous work to maintain privacy have focused on video surveillance only, which is not sufficient for modern multisensor systems. Hence, privacy must be enforced on a higher level of abstraction. In this work, a task- and service-oriented architecture is described, which hosts an Object-Oriented World Model (OOWM) to store information from different sources on a high level. To ensure privacy, access to the OOWMis only granted via a Privacy Manager, Identity Management (IdM) beeing one of its most important functionalities. Recently, privacy and Identity Management have been in the focus of research and many approaches f or e-commerce have been proposed. IdM in the area of surveillance is different and no sufficient solutions exist by now. This work identifies the privacy issues that must be addressed by Identity Management solutions in smart surveillance systems. A new approach for privacy-aware IdM is proposed that copes with the new requirements and shows its integration in a task-oriented surveillance system.