|
Here Below |
Real time speech synchronised animation of a three dimensional facial model
Term : Fall'98 Supervisor: Dr. Prem Kalra The motive was to animate the mouth region of a suitable 3D facial model, on the basis of some input text. The animation model was based on the linear deformation of a wireframe facial model. The text was broken into phonemes, which defined the visemes. Each viseme was mapped to a set of Action Units and represented by 2D lip model with determined the control points. The control points were selected from the set of feature points defined by MPEG-7. These 2D points in (h,ø) space were then mapped on to the 3D vertex points of the mesh. |