This is a simple demonstration of using OpenCV to track the head of the user from the web cam, and adjusting the camera's position and angle accordingly in the 3d environment in Godot; giving a feeling of perspective to the user when they move their head. You can find several examples of this concept on the internet.

The problem with this concept is that detecting the position of the face in a frame, using the CPU can be slow and the user can experience lag when moving their head. Therefore in this code I have used the GPU module of OpenCV which uses CUDA to process each frame using Haar cascades. It is still not perfect, but compared to using the CPU it is significantly better.

This is a module written for the Godot engine, which can be found here.

You need to have a recent version of OpenCV which was compiled with GPU support.

The GPU module of OpenCV is implemented using Nvidia's CUDA so unfortunately you are going to need a CUDA compatible GPU.

I have OpenCV 2.4.10 which I built with CUDA 6.5. I'm using the proprietary Nvidia driver 340.65 on Debian GNU/Linux. I had problems when I tried to build OpenCV with older versions of CUDA so be adviced.

You can find more information about the OpenCV GPU module here.

Copy the cvperspective folder into the modules directory under the Godot engine's root directory. Compile Godot as usual, open the example project included in the example_project directory.

Note: The code uses haar cascades to detect the position of the face. The path to the xml file for the cascade should be specified correctly in the OpenCVPerspective.cpp file.

When the start() function is called in Godot, a new thread is created which connects to the web cam and gets the detected face's position each frame. Using the get_horizontal_offset() and get_vertical_offset() functions, you can get the offset of the user's head from the middle of the screen for the latest processed frame. Using the offset values, you can modify the camera's transform values accordingly.