Anomaly detection in structured/unstructured crowd scenes

Nowadays providing people´s safety in public places is an important issue for governments and security organizations. Anomaly detection in surveillance videos is one of the applications which helps to manage these issues automatically. Particularly, anomaly detection in crowded scenes such as airports, rail stations and etc has attracted a lot of researchers to work in this area, as most of real scenes in surveillance videos are crowded ones. This paper proposes an online approach for automatically detecting and localizing abnormal behaviors in a crowded scene. The crowded scene is modeled from two points of view, a local and a global one. A local confidence factor is introduced which balances between these models based on whether it is a structured crowded scene or an unstructured one. Hence, our algorithm is able to be used for both structured and unstructured crowds in contrast to other algorithms that are only used for structured crowds or unstructured crowds. In addition our approach deals with primary high false alarms in approaches that rely only on local models. Experiments on real-world crowd scenes demonstrate the effectiveness of our approach.