FACE APPEARANCE SYNTHESIS USING FLEXIBLE APPEARANCE MODEL - 3D Face Processing: Modeling, Analysis and Synthesis - page 94

Figure 8.7. Interactive lighting editing by modifying the spherical harmonics coefficients of
the radiance environment map.
2. Face Relighting For Face Recognition in Varying
Lighting
Face recognition has important applications in human computer interaction
and security. Face recognition still remains a challenging problem because
the performance of almost all current face recognition systems is heavily sub-
ject to the variations in the imaging conditions [Phillips et al., 2000], such as
illumination variation.
In Section 1.3.1, we have reviewed approaches that have been proposed to
deal with the illumination effects in face recognition. These approaches need
either multiple training face image per person or 3D face model database for
modeling illumination effects. Zhao and Chellappa [Zhao and R.Chellappa,
2000] tries to use only one image by recovering 3D shape from a single face
image. They use symmetric shape-from-shading but the method may suffer
from the drawbacks of shape-from-shading, such as the assumption of point
lighting sources. Zhang and Samaras [Zhang and Samaras, 2003] propose to
recover the 9 spherical harmonic basis images from a single face image. The
method in [Zhang and Samaras, 2003] needs a 3D database as in [Blanz et al.,
2002] to estimate a statistical model of the spherical harmonic basis images.