Anomaly Detection is system which identifies inappropriate human behavior. One of the major problems in computer vision is identifying inappropriate human behavior. It is crucial as activity detection can help many numbers of applications. It can benefit applications like image monitoring, sign language recognization, object pursue and many more. Many alternatives are there such as low-cost depth sensors, but they do have some drawbacks such as limited indoor use also with lower resolution and clamorous depth information from deep images, it becomes nearly impossible to assess human poses. In order to resolve the above issues, the proposed system plans to utilize neural networks. One of the major research area is to recognize suspicious human behavior in video monitoring, in the field of computer vision. Several surveillance cameras are situated at places like airports, banks, bus station, malls, railway station, colleges, schools, etc to detect suspicious activities such as murder, heist, accidents, etc. It is a tedious job to detect and monitor these activities in crowded places, to trace real time human behavior and classify it into ordinary and unexpected scenarios the system needs to have a smart video surveillance. The experimental results show that the proposed methodology could assuredly detect the unexpected events in the video.
Affective State Based Anomaly Detection in Crowd