A project on pose classification where I use a heuristic approach to count the number of fingers the user is holding. The fingers dataset is drawn directly from the leap motion controller.
The ./fisrt temp/Assets/Script/getInfo.cs contains my heuristic approaches. It works very nice and smoothly :))
Todo list:
- Build a mini FPS game.
- Classify the number of fingers using heuristic methods
- Involve machine learning algorithms for classifying the number of fingers.
- Classify more poses: shooting, defending, go forward, left, right, jump,...
- Deploy machine learning model into the mini FPS game then we have an FPS-hand-controller game without mouse and keyboard.