In may 2023, the mediaipie team released a new API which makes it easy to extract face landmarks from videos and live streams in python and Javascript, which have very clear code examples. This API uses the new tasks to run the updated FaceMeshV2 model. All of this is working well on python and javascript.
I would like to use these models but in C++. Specifically, I would like to run the model FaceMeshV2. But it seems that there is no documentation on how to do this in C++. However, in the code, it seems that python uses the C bindings to do this, using e.g. face_landmarker.h or face_landlmarker.cc.
Does anyone know where I can find clear documentation of how this works on C++? Is there already an open-source project that extracts landmarks with mediapipe in C++? Alternatively, is there an open-source project that runs a mediapipe task in C++? Any info or tutorial on compiling/running mediapipe tasks with the updated API in C++ would be of great help!