Content-based video sequence interpretation

Video sequence interpretation attempts to make sense of events that occur when they have been captured on video. When such scenario understanding can be employed in real-time or near-instantaneously, it can be applied to situations where constant watch would otherwise be needed. The potential for consumer applications is wide-ranging. Examples include safety monitoring of unattended elderly persons, children, pets and security surveillance. In this paper, we present a content-based video sequence interpretation algorithm for confined or semi-confined spaces. The algorithm comprises spatio-temporal video object (VO) segmentation, feature extraction, and scene analysis. Application of the proposed algorithm is illustrated in the context of security monitoring. Our system activates an alarm to alert the user of a suspicious event virtually instantaneously