From range data to animated anatomy-based faces: a model adaptation method

Abstract

This paper presents a new method for reconstructing animated, anatomy-based facial models of individuals from range data with minimal manual intervention. A prototype model with a multi-layer skin-muscle-skull structure serves as the starting point for our method. After the global adaptation, the skin mesh of the prototype model is represented as a dynamic deformable model which is deformed to fit scanned data according to internal force stemming from the elastic properties of the surface and external forces produced from the scanned data points and features. The underlying muscle layer that consists of three types of facial muscles is automatically adapted. According to the adapted skin and muscle structures, a set of automatically generated skull feature points is transformed to drive a volume morphing of the template skull model for skull fitting. The reconstructed model realistically reproduces the shape and features of a specific person and can be animated instantly.