Facial Action Unit mapping to 4D representation

[mno-93]

Hey,

I am wondering if it would be possible to integrate a "facial action unit" (FAU) representation into medusa. This could span across recognition but also rendering.
Facial action units are often used in psychological and avatar research as they're supposed to represent muscle-related movements in the face and seen as the foundation of basic facial expressions.

It would be nice if there could be a mapping of the action units to the 3D mesh and hence, also being able to "manipulate" them and render the effect.

Therefore, it would also be nice to have a "base" mesh that could be adapted by altering specific points and/or action units (over time).