Dominant Flow based Attribute Grouping for Indifferent Movement Detection in Crowd

The growth of technology is heading security systems in a challenging way with amalgamation of different threats and vulnerabilities. Although it is becoming easier day by day to find out the abnormality in the ongoing video with advancement of technology in the field of video cameras, however it is still challenging to detect the undesired event at the time of happening considering the crowd movement in a closed environment. In the following research-paper, we propose a model which proves to be useful and applicable in detection of movement of structured crowd frame by frame in natural sequences. Evaluation results establish the fact that the proposed system is enough capable of detecting the anomalous activities such as merging, splitting or colliding at a point in a certain time than other existing techniques.