Alternative Average Face Models for 3D Face Registration
Neşe Alyüz, Albert Ali Salah, Lale Akarun · 2007
This paper investigates the effect of registration on 3D face recognition, and proposes new methods for average face model based registration. Different approaches for computing average models are contrasted. Two dense registration methods are investigated: Iterative closest point, and thin-plate splines. Our results show that using multiple average face models for registration can increase the accuracy of the classification.