真实感人脸语音同步动画系统

Expressive lip-synchronization facial animation system

提出了一个人脸动画与语音同步的系统,重点解决协同发音、表现人脸细微表情特征的问题。输入带有情绪标志的文本,就能够产生对应表情的、与语音同步的人脸动画;本系统能够生成各种不同性别、年龄、表情特征的高度真实感3D人脸模型,人脸细微表情特征(如额头皱纹)可以随人脸表情的变化而动态改变。基于语言学理论,本系统提出了解决协同发音问题的一套规则。

This paper proposed a system carring out synchronization of facial animation and speech. Solved these problems about coarticulation and modelling facial subtle expression characters. Gave an input string of text annotated with emotion-tags, the proposed system could produce the corresponding lip-synchronized motion of a 3D face model with appropriate expression. This system was able to produce highly real 3D facial models which vary from gender, age, expression and so on. Also facial subtle expression characters, such as forehead furrow, could make dynamically corresponding change with the variety of facial expressions. Based on linguistic theory, proposed a set of rules to solve the problem of coarticulation effect.