Much of what we have covered through Chapter 7 has been about how to interact with the Kinect’s data purely through code. While this grants us an endless capacity to develop very specialized Kinect applications, we risk reinventing the wheel for many common tasks.
Just consider the undertaking of detecting gestures. The most rudimentary manner to approach it would involve some form of heuristics; for example, if the hand joint is higher than the head joint, the person has their hand raised. It is not difficult to see how this increases in complexity for even ...