Animating Faces Using Appearance Models

Here we present an overview of a system we are developing for animating faces using appearance models. To facilitate more widespread use of animated characters, we are interested in developing techniques that: 1) require only easy to use and inexpensive data capture equipment, 2) are threedimensional allowing virtual characters to be animated, 3) allow a range of face models, from near-photoreal to cartoon-like, to be animated depending on the application, 4) allow visual speech information and facial gestures to be easily transferred between models, 5) can be rendered in real-time, and 6) can generate animated sequences directly from motion captured data, lipsynced to audio, or generated entirely from text.