" We can change the world and make it a better place. It is in your hands to make a difference" - Nelson Mandela This is our little effort towards the betterment of society. Public places which deal with a huge crowd, remote places where a regular security person is unavailable, places where security persons are not enough to keep an eye. Keeping all this situation in mind we tried to modify the idea of CCTV cameras, Which are of no use when the crime is taking place in most of the cases. What is the use of those cameras when we refer them after all the destruction has taken place? Our model is to aid the present technology and make it self sufficient to detect crime and alert people on spot.

What it does

It is an application which is combination of 2 model(Suspicious object detection, Facial emotion recognition) It will predict the Criminal Activity using these model however we'll add one more model(Pose estimation) to increase the accuracy. if a person is carrying Gun,Knife etc., Suspicious object it will detect that and alert the concerned person and only Object if Person is angry then also it will alert and with the help of Behavioral of that person(Pose estimation) It's accuracy can be increased.

How we built it

We used deep neural network, OpenCV and Keras to accurately detect the Criminal Activity and alert the concerned person. A combination of stereo-camera and machine learning algorithms was utilized to create an accurate depth map of the video stream. Whole code was written in Python.

Challenges we ran into

The main challenge we ran into was pose estimation which is used to detect criminal activity via detecting the suspicious behavioral activity of that person

What's next for Criminal Activity Prediction

In Future we would like to integrate Pose estimation and many more model algorithm to increase the prediction of Criminal Activity.

Built With

Share this project: