Hand Cut uses hand gestures to perform shortcuts.
The program uses camera-based hand tracking to detect your hand's gesture.
Made with Python using OpenCV and
Mediapipe.
The idea is an extension of one of my projects, Squat King. With more of a focus on
customizability, since you can asign any pose with a shortcut. Thus meaning I need a way to save and store landmarks, for which I had
to create custom JSON parser and serializer.
I also needed a way to compare a "stored" pose with the pose that's being executed,
which opens a lot of decisions (should the pose take into account the position on-camera? does orientation matter? which is the main landmark? what is the error margin? etc.)
Overall it was pretty easy to make since I had already done Squat King,
nontheless I learnt many new things and expanded my knowledge on old ones.
It was a little experiment so it lacks QoL and basic functionalities.
I might (or might not) update it in the future by adding either a menu with Tkinter
or a small CLI.