Analysis and Synthesis of Facial Expressions - 3D Modeling and Animation: Synthesis and Analysis Techniques for the Human Body - page 256

Figure 8: Animation of different people using facial expressions from a
reference sequence. Upper row: Reference sequence; Middle and lower
row: Synthesized new sequences.
a texture map. Since the topology of the mesh is identical for all models, the
surface deformation description need not be changed and facial expressions can
easily be applied to different people.
Since the same generic model is used for all people, point correspondences
between surface points and texture coordinates are inherently established. This
enables the morphing between different characters by linearly blending between
the texture map and the position of the vertices. In contrast to 2-D approaches
(Liu et al., 2001), this might be done during a video sequence due to use of a 3-D
model. Local deformations caused by facial expressions are not affected by this
morphing. Figure 9 shows an example of a view of the morphing process between
two different people.