Ontology of human identification by face and body motions in video surveillance systems

Abstract

At the present stage of advancing information technology, the development of models and recognition methods by body movements and faces in video surveillance systems is a topical problem. This task is essential for security issues, especially at facilities with mass gatherings to counter a terrorism-related crime. The paper presents a classification of the main biometric features and parameters that characterize a potential violator. This classification has been developed for security control systems and access systems of enterprises. A block diagram of merging biometric data and violator recognition by body motions and face which can be used as the basis for the development of security control systems is proposed. The types of systems and methods of human recognition by body movements and face are considered, their advantages and disadvantages are revealed. It is noted that for accurate violator recognition under a set of biometric features, it is reasonable to use a combination of recognition methods which will allow to make the right decisions regarding the identification of a potential violator. This paper attempts to consider the main aspects related to human recognition by body movements and face in video surveillance in general, in contrast to well-known works devoted to individual biometric features.