I'm writing a simple game with a Kinect camera. However, I'm unable to figure out how to check for collisions. The objects are rectangles and a depth map from Kinect (which will be a human-like shape). I'm using SDL for the graphics.
I've thought and searched for days but not found a good answer and I'm running out of ideas.
My main direction has been to colour the outline of the rectangle in one colour and the depth map in another. Then in every frame the program checks if the outline has got the colour that it is supposed to. If that's not the case, it means that the depth map is over the box and it's a collision. The problem is here that I've so far only found examples that reads the pixels of a texture, not the whole screen.
Am I at the wrong path? Or am I missing something easy?