Hand Gesture Recognition System Using Mediapipe and Tensorflow

I would like to show you one of my research project on hand gesture recognition system. Actully the idea behind this project is to make a gesture recognition system that can be integrate in different applications and use cases. The possible use cases can be:

  1. Controlling the smart home devices just by hand gestures.
  2. Controlling the computer using by just hand gestures.
  3. Controlling vehicle operations such as virtual driving steering, dashboard controls just by hand gestures.
    This codes are implemented by using the Tensorflow and Mediapipe models to make it more accurate and simplified. Although the implemented codes are just a start and I am working on the advanced version to see in different application areas.
    The source codes are available here: GitHub - rohitnarain24/hand_gesture_recongition_mediapipe
    The demo of system is avaliable here: Hand_gesture_recogniton - Google Drive

I hope this will help the others who wants to learn how hand gesture recognition system can be implemented.

1 Like

Interesting idea Rohit,

How was your experience using Mediapipe?
any feedback on the library?

1 Like

Thanks Igsum, for your kind words!
Yeah it was a amazing experience with the library. Although I feel some drop in accuracy in real time like for some gestures the model was seems like confusing. But I believe we can add some more patterns into the training.
But overall… Its a amazing experience for me… Thanks for this great library

1 Like