Because I have not been able to get any other facial rendering code until very late into the project, I have based the rendering in my project on Parke's model. This code, which is based on affine transformations on a mesh of points, has its limitations, and is not as realistic as more recent facial rendering code (such as Waters' [4] system) which uses more sophisticated physical models. Possible further work could include the adaptation of this code to use a more modern rendering technique. This would allow other actions (such as puffing of the cheek) to be rendered.