A socio-technical approach for event detection in security critical infrastructure

Philipp Blauensteiner, Martin Kampel, Christoph Musik, Stefan Vogtenhuber

Abstract: In recent years, surveillance and monitoring have become topics of increased interest for not only research active in computer vision, but also for many scholars in the field of social sciences and the humanities. Joint research becomes important when effective surveillance systems are deployed in public spaces. This paper focuses on designing and developing a socio-technical surveillance application in a security critical infrastructure like a bank. The technical system presented is designed to detect specific events like abnormal or suspicious behaviour. A close collaboration between scientists from Computer Science and Sociology actively reflects the development of the new surveillance technology. The novelties of the paper are rather a socially intelligent surveillance system coordinating human and non human actions then technical achievements. Based on interviews several behaviours of interest as well as an interaction between the system and its users are defined. Experiments and results are shown based on a real life installation.