Global Abnormal Behaviour Detection Using a Network of CCTV Cameras

This paper investigates the detection of global abnormal behaviours across a network of CCTV cameras. Although the problem of multiple camera tracking has attracted much attention recently, little work has been done on modelling global behaviours of objects monitored by a network of CCTV cameras with disjointed camera views, and no effort has been taken to tackle the challenging problem of detecting abnormal global behaviours, which are only meaningful and recognisable when observed over space and time across multiple camera views. To that end, we propose a novel framework, which consists of object tracking across camera views, global behaviour modelling based on unsupervised learning, and probabilistic abnormality inference. The effectiveness of the framework is demonstrated with experiments on real-world surveillance video data.