Expressive Facial Animation Based On Visual Feature Extraction

As the important subject in computer animation, human-computer interaction and computer vision, the facial animation provides us great convenience and fun on communication, culture and entertainment. As a matter of fact, facial animation is widely applied in the movie/TV, visual teaching, network conference and video-games.The rapid and important progress is being made on the research of facial animation, which involves a lot of disciplines such as physiology, psychology, art, computer graphics and computer vision. Many researchers dedicate themselves in this area and provide us more and more impressive approaches on facial animation. Nowadays, the state-of-art techniques can bring us very expressive, stylized facial animation. The facial animation can be also driven by speech or text, which greatly broadens the applications.In this paper, we firstly extract the facial shape feature from the video clip. And then, an expressive facial animation approach is developed based on the input video and the recognized feature. This approach contains two steps. Step one: a key-frame generation method which uses few recorded frames to create the whole key-frame sets containing all the typical facial variation and expression. Step two: a fast, incremental intermediate-frame generation algorithm, which inserts intermediate-frame between key-frames based on the key-frame sets and the input parameters. Finally, the smooth and impressive facial animation is obtained.Our experiments show that the approach in this paper can build facial animation with rich expression, which can be also easily combined with voice-driven technique. At the same time, this approach can make cartoon-style facial animation.