A framework for gesture recognition using machine learning.
In order to use this framework, please install
- Android Studio
- python3
- matplotlib
- numpy
- pandas
- tensorflow
The DataRecorder app records movement using the smartphone's sensors. With the DesktopApp, you can automatically build the model. The ModelTester app is used to test the models' capability to classify movement.
The idea for this project arose from a uni course. Here, the goal was to dynamically adapt the user interface of a note-taking app to the user's hand. You can read this article, if you're intersted. This also shows possible use cases for such a kind of framework.
I don't do any maintenance for this project anymore because I don't have time for it at the moment. This is also an older project, so the code standard is not really good yet 😅, but please feel free to be inspired by this project.