Human face modeling by fitting a 3D anatomy-based model using dynamic deformable meshes

Abstract

This paper presents a new adaptation-based method for reconstructing animatable human facial models from incomplete scan data with minimum user intervention. An anatomy-based generic control model serves as the starting point for our reconstruction algorithm. A global alignment automatically adapts the position, size, and orientation of the generic model to align it with the scan data based on the measurements between a set of sparse anthropometric landmarks. In the subsequent physically-based face shape adaptation, the generic model is represented as a dynamic deformable mesh which subjects to internal force stemming from the elastic properties of the surface and external forces generated by input data points and features. The adaptation is governed by equations of motion which are it-eratively simulated to recover personalized face shape. The reconstructed 3D face resembles the shape of the individual and can be animated to synthesize various facial expressions with transferred virtual muscles.