Tensor-based 3D color face reconstruction using a single image

Abstract

This paper develops a tensor-based 3D face reconstruction approach from a single color image. Instead of considering grayscale images, this paper also takes additional color factors into account while constructing the tensor model. Canonical correlation analysis then is applied to establish the relationship between the color image and the depth information in the face database. During the face reconstruction, given a single color face image, the depth estimation is recovered from the CCA-based mapping between the tensor models. Experiments show that the proposed approach improves the reconstructed results under different lighting conditions and poses.