It is a deep learning model that takes videos as input and divides them into frames and then passes it to the model having two autoencoders generating scores and then from those scores generating abnormal frames
The budget is negotiable and the max time limit 8-10 days