Physical facial model based on 3D-CT data for facial image analysis and synthesis

Abstract

The authors have proposed a physics-based facial model based on anatomical data and knowledge, which has a three dimensional structure for facial image analysis and synthesis. The skin and the skull models are constructed from 3D-CT scanned data, and the facial muscles were placed between them just like as a real human face. With regard to the skin and muscle models, spring frames are used to simulate the elastic dynamics of a real facial skin and muscles. Facial modification can be flexibly realized by two factors: the expansion and contraction of non-linear springs which simulate a skin and muscles, and the motion of the jaw part of the skull. Using the system, one can represent realistic facial animation and predict facial modifications under given conditions for muscles and the skull. Furthermore, the method is effective for analyzing the relationship between modification on the facial surface and the motion of muscles and the skull.