3D Face Reconstruction from a Single Image Using a Deformable Generic Elastic Model Robust to Changes of Facial Expressions

Abstract

In this paper, a new and efficient method was proposed to reconstructing 3D models of a human face from a single 2D face image robustness under a variety of facial expressions using the Deformable Generic Elastic Model (D-GEM). The Generic Elastic Model (GEM) approach was extended and combined with statistical information of the human face and deformed generic depth models by computing the distance around face lips. Particularly, it was demonstrated that D-GEM could estimate the 3D model of the frontal face image more precisely and achieve a better and higher quality of 3D face modeling and reconstruction robustness under a variety of facial expressions compared to the original GEM and Gender and Ethnicity-GEM (GE-GEM) approach. It was tested on an available 3D face database, which demonstrated its accuracy and robustness compared to the GEM and GE-GEM approaches under a diversity of imaging situations including facial expressions, gender and ethnicity.