Graphics Reference
In-Depth Information
Figure 8.74
Rendered images of reconstructed hair geometry (right), giving a “ground truth” compari-
son to captured images (left). (From [Jakob et al. 09]
c
2009 ACM, Inc. Included here by
permission.) (See Color Plate XIX.)
scheme to extrapolate already existing hair directions to the empty regions. The
method is only briefly sketched in the paper, and the authors point out that in the
case of a dense hair cluster (such as tightly braided hair) other approaches may
be more suitable. Specifically, the infill generation method in the paper “Hair
Photobooth” [Paris et al. 08] could be employed instead.
Figure 8.74 compares rendered images of the reconstructed model (right) to
actual captured images (left). The images were rendered using the particle simu-
lation approach of Moon et al. described in Section 8.5.1. The accuracy of the re-
covered geometry is remarkable. The “fiber by fiber” approach is a major advance
in capturing fine detail of hair, and clearly shows the importance of fiber-level de-
tail in realistic hair rendering. Data produced by this method can be expected to
impact hair rendering technology in the immediate future.
8.6.2 Acquisition of Hair Color
A method to better reproduce the apparent color of real hair was proposed in
the paper “A Practical Approach for Photometric Acquisition of Hair Color” by
Arno Zinke, Martin Rump, Tomas Lay, Andreas Weber, Anton Andriyenko, and
Reinhard Klein [Zinke et al. 09]. The goal of the work presented in this paper
