Robust 3D face shape estimation using multiple deformable models

In this paper, we present the robust and fast 3D face shape estimation system which consists of two mirrors and a camera. In the face shape estimation procedure, we propose the iterative shape estimation algorithm using the multiple deformable models. In the experiments, we reduce the processing time by using only few vertices in an input face image. Our proposed algorithm shows accurate face shape estimation result when compared with the ground truth. To evaluate the robustness to the positional errors in our algorithm, two experiments are carried out. The results show that our proposed algorithm is more robust to the positional errors than other 3D reconstruction algorithms.