This paper addresses the problem of adapting a generic 3D face model to a human face of which the frontal and profile views are given. Assuming that a set of feature points have been detected on both views the adaptation procedure initializes with a rigid transformation of the model aiming to minimize the distances of the 3D model feature nodes from the calculated 3D coordinates of the 2D feature points. Then, a non-rigid transformation ensures that the feature nodes are displaced optimally close to their exact calculated positions, dragging their neighbors in a way that does not deform the facial model in an unnatural way.
Nikolaos Sarris, Michael G. Strintzis