A deep learning framework to identify abnormal activity in closed space.

In today's world its getting common to have security breaches in closed and restricted activity areas like ATM's and banks. this framework has been proposed to get over with these problems. with this proposed framework the system will be able to identify the face as well as the full body parts of such intruder or abnormal actor to say. we have used open cv and other 3 models to work on our framework. we have used MTCNN (multi-task cascade convolutional neural network) as one of our models instead of the basic CNN. [ABSTRACT FROM AUTHOR]