Although direct manipulation for figure animation has long been possible using inverse kinematics, there has been no similar "inverse kinematics" approach for the popular blendshape facial representation. In both problems the system must infer unknown degrees of freedom during each edit. We solve this inverse problem for faces by minimizing changes in facial expression. The new approach is obvious, easy to implement, and interoperates with traditional blendshape parameter (slider) editing, unlike face editing approaches based on an underlying PCA representation. This is crucial, since a simple mathematical argument shows that direct manipulation is sometimes less efficient that parameter editing, while the converse is equally true in other cases.

J.P. Lewis and Ken Anjyo
Direct Manipulation Blendshapes
IEEE Computer Graphics and Applications Special Issue: Digital Human Faces
30(4), 2010, p.42-50.

Video (5M .mov)