A Real-time suspicious behavior detection system using deep learning

Abstract

Big data applications are consuming most of the space in industry and research area. Among the widespread examples of big data, the role of video streams from surveillance cameras is equally important as other sources. Surveillance videos have a major contribution in unstructured big data. Surveillance cameras are installed in all places where security is crucial. Security can have different meanings in different contexts, it covers all types of abnormal activities including theft identification, violence detection, terrorism ..etc. Manual surveillance is tedious and time consuming. Throughout this research we introduce computer vision and deep learning techniques as an alternative way of dealing with video feed analytic, as detecting suspicious behavior and making the process more automated and reliable.