Home People Projects Teaching Publications Contact Jobs
 
+    
Computer Aided Face Engineering
   
 

       

Recently we introduced the concept of correspondence based face space [1,2,5]. The technique derives a general flexible face model, which allows to synthesize novel realistic 3D-models of human faces [7]. The flexible model is obtained in an automated learning procedure from the example faces of our 3D face database [6]. The procedure extracts the common structure of all faces, the prototype face, and codes individual faces through their deviation from this prototype. To develop this model, which is mainly based on the linear combination of examples, we had to solve the correspondence problem beteen the indiviual example faces. That is, we had to develop algorithms that are able to find all corresponding points on the surface of the example faces.
       
 

   

Applications:

Face Synthesis by linear Combination
 

    3D-Model from 2D-Image

 
    3D-Caricatures

 
              Face Expressions
           
 
              Manipulation of Sex
           
     
  References:
[1] Synthestis of novel views from a single face image.
    T. Vetter, International Journal of Computer Vision, 28:2 (1998) 103-116.
[2] Linear Object Classes and Image Synthesis from a Single Example Image.
     T. Vetter and T. Poggio, (1997) IEEE PAMI, 19(7):733-742.
[3] Three-dimensional caricatures of human heads: Distinctiveness and the perception of facial age.
     A.J. O'Toole, T. Vetter, H. Volz and E.M. Salter, Perception, 26 (1997) 719-732.
[4] Sex classification is better with three-dimensional head structure than with texture.
     A.J. O'Toole, T. Vetter, N.F. Toje and H.H. Bülthoff, H. H. (1997) Perception, 26:75-84.
[5] Separation of texture and shape in images of faces for image coding and synthesis.
      T. Vetter and N.F. Troje, Journal of the Optical Society of America, A 14:9 (1997) 2152-2161.
[6] A bootstrapping algorithm for learning linear models of object classes.
    
  T. Vetter, M.J. Jones and T. Poggio, (1997) IEEE CVPR'97, pp 40-46, Puerto Rico, USA.
[7] A Morphable Model for the Synthesis of 3D Faces.
       Blanz, V. and Vetter, T. (1998)   SIGGRAPH'99 Conference Proceedings pp. 187-194