ABSTRACT
A facial rig for facial motion captured data can be set up in a number of ways. The simplest way is probably attaching markers to a set of joints that are bound to the facial skin geometry. Another approach is to create “muscles” that are polygons placed under the facial skin around the mouth and eyes and on cheeks, forehead, chin, and any other part of the face that needs to be animated by mocap data. The first step of setting up a facial rig with inverse kinematics is creating "skin" joints and "muscle" joints. One way to get rid of the head movement is to not allow the captured subject to move the head in the first place. The Facial Data Stabilizer considers head motion as the rotation and translation of an object in the world space, while the deformation of facial features is considered as translation of the object's vertices in the object (local) space.
