“3D facial animation from high speed video” by Suzuki, Ishibashi, Kubo, Maejima and Morishima
Conference:
Type(s):
Title:
- 3D facial animation from high speed video
Presenter(s)/Author(s):
Abstract:
Synthesizing facial animation of real individuals is a time-consuming task. In this paper, we present a method of generating facial animation automatically with 2D high speed video. First, we create a 3D personal face model by measuring 3D range data from an actor’s neutral face. We then capture 2D high speed (super slow motion) video which includes target changes in the subject’s facial expressions in natural situations. After tracking feature points on the face in video sequences, all of the vertex positions of the 3D face model can be decided using Radial Basis Functions (RBF) [Arad, Reisfeld. 1995]. Additionally, the texture of the 3D face model is updated automatically frame-by-frame using the video images.
References:
1. J-Y. Bouguet, Pyramidal implementation of the Lucas Kanade feature tracker description of the algorithm, Intel Corporation, Microprocessor Research Labs, 1999.
2. N. Arad, D. Reisfeld, Image Warping Using few Anchor Points and Radial Functions, The Eurographics Association 1994.
