To detect gestures, you have to do a lot of work. Recognizing hand movements and turning them into events for the application is not a trivial task. In order for the application to recognize a hand wave, for instance, you can take two approaches:
- Procedural: This is done in code. You detect a hand position, note its placement in 3D space, then see the movement and deduce from that what happens.
- Machine learning: This is a way to use the power of computing to learn about gestures. Basically, you feed an enormous number of images to a computer and try to teach it to recognize the gesture you want it to see.
Both approaches have advantages and disadvantages. For instance, the procedural way is relatively easy to get started with, but usually the results are less accurate. The machine learning way is hard to set up and takes a long time and a lot of training...
