Face Model Reconstruction via Camera Pose Estimation Under Orthographic Projection

Abstract

In this paper, we developed a systematic approach for the reconstruction of a 3D mesh model of a human face from multiple images in a semi-automatic way. Our approach can be divided into three steps: (1) estimation of the camera poses from multiple images under the assumption that the camera projection model is orthographic, (2) computation of the 3D coordinates of a set of control points with multi-view stereo, and (3) reconstruction of a mesh model of human face by deforming a generic model to be fit with the control points. Experiments results show that our method is effective for reconstruction of human face models.