Approximation of 3D Face Model
Taher Khadhraoui, Faouzi Benzarti, Hamid Amiri · International Review on Computers and Software (IRECOS) · 2013
The approximation of a 3D face model is considered very promising in the recognition of individuals. However, the challenges are open today: the robustness of these 3D approaches to change facial expressions and optimization of computation time required in the scenarios identified. In this paper, we present a complete solution for the extraction of a 3D model of face and its 3D position from a small number of pairs of landmarks 2D-2D and 2D-3D matching. This extraction step is considered as an initialization method for face tracking based on the 3D model. The solution is divided in two steps. The first step recovers an approximation of the 3D face model, the second step deals with the improvement of this 3D face model and the pose extraction. Our solution is robust, fast and sufficiently descriptive.