Description

Several humans are tracked in videos obtained from several cameras. Detection is the initiation of a subject track and it is performed independently in each camera view. An initial state vector is established for subsequent tracking. We propose sequential importance resampling particle filter(SIRPF) to track a human in a video. This reduces the computing complexity and produces accurate results. The tracking nesserey for find out the abnormal event in video tracking. After the tracking we find the entropy values each moving partical and find the partical moving varation by using the gmm classifers that movement variation is identify the abnormal event in crowd scenario.