Abstract:

The authors adopted the combination of weightd Dirichlet freeform deformation algorithm and thought of gradient animation to build the textdriven and speechdriven facial animation system. It has solved many problems in traditional methods such as the complexity of modeling process, the low fluency in animation output, the inconsistency between lip and voice and so on. Because of these advantages it achieves a more realistic realtime animation and shows very positive practical and theoretical value as reference.

Key words: voice recognition, facial animation, voicedriven, textdriven, Dirichlet freeform deformation, gradient animation