Generation of 3-D head models from multiple images using ellipsoid approximation for the rear part

A system for building a three dimensional (3-D) human head model from two camera views is proposed. A generic 3-D face model is adapted to a human face of which the frontal and profile views are given. Then, a 3-D ellipsoid approximation technique is used to fit an ellipsoid to the 3-D wireframe model. This ellipsoid is then used to approximate the back of the head, while the the initial 3-D face model provides information for the high-detail front part of the head. The part of the ellipsoid model that corresponds to the back of the head is then identified and an algorithm to merge the front (face) and the back part is proposed. A cylindrical texture map is finally built covering the entire area of the head by exploiting the inherent face symmetry. The final result is a complete, textured model of a specific person´s head