Facial feature localization
The following is experiment result of locating
facial features in AR face database.
The first column is the mean shape. The second column is the result of AAM. The
third column is the result of Shape Inference. The fourth column is the result
of RankBoost. The fifth column is the result of our algorithm. The ground truths are the red
dots and the mean shape and the localization results are green dots. The
localization errors are shown at the top of the images. The sizes of the following images are the half of the size used in the
experiment.
Refer to Section 5.3 for experiment details. Refer to Table 1 (C) and Figure 8 (C) for the benchmark results.
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |