Anomaly Detection in Video Surveillance using SlowFast Resnet-50

Surveillance systems are widely used in malls, colleges, schools, shopping centers, airports, etc. This could be due to the increasing crime rate daily life. It is a very tedious task monitor and detect abnormal activities 24x7 from surveillance system. So detection of events videos hugely demanding area research. In this paper, proposed framework for deep learning concepts. Here SlowFast Resnet50 has been extract process features. After that, neural network applied generate class using Softmax function. The UCF-Crime dataset Graphics Processing Unit (GPU). includes 1900 with 13 classes. Our algorithm evaluated by accuracy. works better than existing algorithm. achieves 47.8% more accuracy state art method also good compared other approaches detecting activity on dataset.