Graphics Reference
In-Depth Information
Figure 1.9 Two different frames out of a 1000-frame face video sequence Hernandez et al. (2011).
The left column shows the reconstruction when shadows are ignored. Middle and right columns show
the corresponding reconstructions after detecting and compensating for the shadow regions using the
shading regularization scheme (middle) and shape regularization scheme (right). Note the improvement
in the regions around the nose reconstruction where strong cast shadows appear (see arrows). Note also
how the shape regularization scheme fails to reconstruct some boundary regions (see circle). Copyright
C
2011, IEEE
Figure 1.10 Face sequence. Acquisition of 3D facial expressions based on Hernandez et al. (2007) and
the shadow processing technique described in Hernandez et al. (2011). The shadows are processed with
the shading regularization scheme. The full video sequence has more than a 1000 frames reconstructed.
Copyright C
2011, Springer
Search WWH ::




Custom Search