3D Face Reconstruction Based on Geometric Transformation

Abstract

Most of current 3D face reconstruction methods based on 2D image are complex. In this paper, a novel and efficient method on the basis of geometric transformation is presented. The method is composed of two aspects, pose estimation and 3D model reconstruction. For the first part, an improved Active Shape Model(ASM) algorithm is used to detect the feature points of the face in the image, and the frontal face information is got based on the symmetry of human face. For the second part, first, establish a correspondence between the key facial feature points and the generic model points, then divide the face model into different regions and adjust them separately based on affine transformation. Experimental results show that the proposed algorithm provides satisfactory results.