Simple Python wrapper for Google's Mediapipe Hand Tracking pipeline.
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.
These are required to use the HandTracker module
numpy opencv tensorflow
To run the Jupyter demo you'll also need
Download models using load_models.sh, or load them manually from metalwhale's repo and put them inside models dir.
To get the SSD anchors I've written a C++ program that executes the
SsdAnchorsCalculator::GenerateAnchors function from this calculator.
As there's no reason to modify provided anchors I do not include it into the repository, but you can find the script here
This work is a study of models developed by Google and distributed as a part of the Mediapipe framework.
Big thanks to @metalwhale for removing cusom operation dependencies.