The challenging task is to identify the human activity by the CCTVs. Every person cannot be monitored in this growing era. There are cameras installed on every step in the smart cities. Over 8,000 cameras are installed in the city Lahore to monitor the suspicious activities and supervision. Human surveillance over these cameras requires great effort.　
For this purpose, machine can be made intelligent with the help of some neural network. Cameras can detect any type of suspicious activities by their own. If any of the suspicious activity will happen under the range of camera, it will be detected and pass on to the buzzer for further action. The suspicious activity can be recognized as snatching of mobile phones, purse and other things. The snatching which is done on gun point will also be recognized by the camera, murder of someone and the fighting scene would also be detected by the camera. This idea can be planted with the help of drones for the surveillance of massive crowds. To keep an eye on the massive crowd and its suspicious activity can be detected with the help of this model. The idea can be helpful in the streets of the city to reduce the crime.