Many techniques have been proposed to acquire 3D shape from 2D images. Photogrammetry of stereo images can be used to determine the shape of faces. Two images of an object are taken from different viewpoints. Corresponding points are found between two images, and the 3D coordinates of the points are determined by triangulation. Automatically finding corresponding points is a difficult problem so meshes may be drawn on the faces. The first 3D facial models were created using this technique [105][103].
The 3D shape of faces can be determined from a single 2D image of it, if a regular structure (such as a line array or a square grid) is projected on it [150].
Another technique is that a generic facial model is prepared and transformed to a specific facial model that is consistent with photographs of an individual face [75]. The advantages are that no meshes need to be drawn on faces and that no special hardware is required. The drawback is that it is basically impossible to acquire the detailed structure.