A Framework for Anomaly Event Detection by Analysing the Video Sequences

Currently, it is very essential to monitor activities in Video surveillance applications both in private and public environments. In this context, our paper presents a novel framework to detect the normal or abnormal event situations by analyzing the pixel-wise motion context using block-based approaches. First motion estimation techniques are applied to characterize the events at the pixel level. Optical flow is used to extract information such as density and velocity of motion. The two different proposed approaches identifies abnormal motion variations in regions of motion activity based on the entropy of Discrete Wavelet Transform and Discrete Cosine Transform coefficients. The successful results of the detection of normal or abnormal events on different datasets are reported.