Modeling

Most facial modeling systems describe facial actions with either muscle notation or FACS (Facial Action Coding System) [37]. FACS is based on anatomical studies and denotes any visible movement.

• Libraries of expression and interpolation: In early models, modeling was done by digitizing sculptures of the face with various expressions (different lip shapes and expressions) and storing them in a library [73][41]. Animation was performed by linear interpolation between given stored expressions. Such a method is really tedious and time consuming since it is not automatically adaptable to any other new model.

• Parametric models: Some other systems [108][88][105] were based on a set of parameters which affect not only the size and structure of the model (long nose, short forehead) but also its expressions (opening of the mouth, raising eyebrows). The separation between conformation parameters and expression parameters forces the independence between facial features and the production of an expression.

• Physically-based models: These systems describe the skin as an elastic spring mesh where unit actions are simulated by forces. The deformations are then performed by solving the dynamic equations [144][123]. Muscle movement propagation is intrinsic to the model. Various layers of facial tissue are integrated [135]. It succeeds in producing subtle facial actions with realism.

• Integration of additional features: Texture mapping and hair modeling have been used to enhance realism of the model [2][75][151]. The consideration of wrinkles and of aging effects [138] adds much to the rendering of facial skin and expressions.