Abnormal Event Detection in Video using Appearance and Motion Information

This paper presents associate approach for the detection and localization of abnormal events in pedestrian areas. The goal is to style a model to observe abnormal events in video sequences exploitation motion and look info. Motion info is described through the employment of the speed and acceleration of optical flow and also the look info is described by texture and optical flow gradient. Not like literature ways, our planned approach provides a general resolution to observe each international and native abnormal events. What is more, within the detection stage, we have a tendency to propose a classification by native regions.