The feature coordinates (green dots) are from 3D reconstruction of two pictures.
Blue dots control and deform the head model by radial basis function.
RBF code/tutorial is available from http://cg.alexandra.dk/
--------------------------------
3D feature positions of head are obtained using multiple view geometry. The concept is not overly difficult, but I also cannot explain it with just few words. I'll give you some links instead below. A few days of studying with the second link should get you good enough!
http://en.wikipedia.org/wiki/Epipolar_geometry (overall concept)
http://users.cecs.anu.edu.au/~luke/cvcourse.htm (under 3D Computer Vision)
Some manual labors are required to get the points such as camera calibration, manually picking feature points from two images, etc.
Looking good. How did you determine the 3D position of the feature points based on the 2 images?
Pascal123M 11 months ago