Reconstruction of 3D faces by shape estimation and texture interpolation

Abstract

This paper aims to address the ill-posed problem of reconstructing 3D faces from single 2D face images. An extended Tikhonov regularization method is connected with the standard 3D morphable model in order to reconstruct the 3D face shapes from a small set of 2D facial points. Further, by interpolating the input 2D texture with the model texture and warping the interpolated texture to the reconstructed face shapes, 3D face reconstruction is achieved. For the texture warping, the 2D face deformation has been learned from the model texture using a set of facial landmarks. Our experimental results justify the robustness of the proposed approach with respect to the reconstruction of realistic 3D face shapes.