Suspicious Action Detection in Intelligent Surveillance System Using Action Attribute Modelling

Research in the field of image processing and computer vision for recognition suspicious activity is growing actively. Surveillance systems play a key role monitoring sensitive places such as airports, railway stations, shopping complexes, roads, parking areas, banks. For human it very difficult to monitor surveillance videos continually, therefore smart intelligent system required that can do real time all activities categories between usual some abnormal activities. In this paper many different has been discussed. More focuses given violence like hitting, slapping, punching etc. large action dataset UCF101, Kaggel required. This proposes method model actions using Gaussian Mixture Model with Universal Attribute Model. vector used remove redundant attributes get low dimensional relevant vectors.