Facial Expression Invariant 3D Face Reconstruction from a Single Image Using Deformable Generic Elastic Models

In this paper, we propose an efficient method toreconstructing the 3D models of a human face from a single 2Dface image robustness under a variety facial expressions usingthe Deformable Generic Elastic Model (D-GEM). We extendedthe Generic Elastic Model (GEM) approach and combined itwith statistical information of the human face and deformedgeneric depth models by computing the distance around face lips.Particularly, we demonstrate that D-GEM can approximate the3D shape of the input face image more accurately, achieving abetter and higher quality of 3D face modeling and reconstructionrobustness under a variety of facial expressions compared to theoriginal GEM and Gender and Ethnicity-GEM (GE-GEM)approach. It has been tested on an available 3D face database,demonstrating its accuracy and robustness compared to theGEM and GE-GEM approach under a variety of imagingconditions, including facial expressions, gender and ethnicity.

Keywords:

3D modeling , deformable models , face reconstruction , generic elastic models