Abstract: Blendshape models are commonly used to track and re-target facial expressions to virtual avatars using RGB-D cameras and without using any facial marker. When using blendshape models, the ...