Context-based Authorization in Video Surveillance

The use of video surveillance in the developed world has significantly increased over past two decades, due to decreasing hardware costs and the sense of insecurity ignited by incidents like 9/11 and London bombing. Compared to traditional video surveillance systems which used dedicated transport channels and were observed in a closed monitoring room; modern systems offer advanced functionalities such as object-detection and tracking an individual. Having such systems in place raises privacy issues as the video streams can now also be watched by the patrolling guards over their hand-held devices through Internet. Addressing privacy concerns requires measures such as taking contextual information of the observers into account while allowing access. One of the important contextual information is location of observer, to ensure that observer is located within allowed premises. The aim of the project is to devise access control framework that authorizes individuals to see video streams considering their positional information. Doing so requires gathering user's location through GPS, for example, and the coordinates are mapped to the geospatial model of the allowed area. For mapping geospatial information, project may use GIS toolkit while GeoXACML may be used to enforce access control policies. Devising the framework includes formulating a context-aware access control model, proposing an architecture to enforce such model and development of a proof-of-concept prototype.