Quote:
Originally Posted by JudgeSmails 
|
I can speak to the RGB camera, in computer vision, your algorithm would of course be very sensitive to things like contrast and lighting which can be affected by things like skin color. More than a skin color problem, this probably highlights how sensitive the device is to these conditions, meaning that it may have lots of trouble if you are playing it against certain backdrops or light levels.
I don't know much about the depth sensor, I wouldn't think it would be affected by this too much. Since the system uses a camera and a depth sensor, you can imagine that at the very least broad motions are detected by the depth sensor while the camera can pick out more details (facial expressions, hand gestures, etc). Again, I would guess the problems pointed in that article have to do more with the more fine grained movements not the broad ones.