CamSpace [6], a vision-based project, can use almost any object for input during the game, using a computer-vision algorithm.
For its work, a pre-step calibration and recognition phase is necessary in order to configure the framework with the object that will be used for the game.
Unfortunately, as with many computer-vision-based approaches, the player also needs to be in the focus of the camera, which could be a problem in cases where many people are going to play the game simultaneously.
In this case, the algorithm can get lost during the game or take too much time for processing data.
Another disadvantage of this approach is the dependence on a camera to work properly.