Craniofacial Reconstruction Based on MLS Deformation

Craniofacial reconstruction aims at estimating the facial outlook associated to an unknown specimen. Craniofacial reconstruction is generally based on the statistical tissue thickness on anthropometric landmarks. However, the features points alone are not enough for realistic reconstruction. So, in our paper, we take advantage of a reference facial model, through measuring the differences between the target facial feature points and the reference facial feature points, a novel craniofacial reconstruction algorithm based on moving least squares deformation is presented. The 3D skull mesh model is obtained with Marching Cube Algorithm, which extract the iso-surfaces from a complete head CT slices datum. The holes detected in the 3D skull model can be repaired with different methods after the holes are clarified. Then, the craniofacial is reconstructed using MLS deformation with the constraint of a reference facial model. The experimental results show that the methods can produce more desirable results than others’. Key-Words: Craniofacial reconstruction, Deformation, MLS, Feature points, Hole repairing