The three top pictures show the result of a project I have worked on
during my education. It is a 3D face model construction tool. It takes
two images of the user, some feature points are placed (manually right
now) and you have a 3D model of your face as a result.
[Special thanks to the sitter, a classmate]
It can be compared to the IOTD by Wolfgang Engel. His program gave me this project idea. But it
does not work in the same way. I am using the 2 pictures as two
different textures, whereas Wolfgang was blending them together in order
to use environment mapping. The two textures are mapped with 2 planar
projection (front and side), and I use an Alpha blending function that
fakes the curve of the face to blend them together.
The four bottom pictures show results from my 3D engine: it can load
Quake I, Quake III levels (ok, who hasn't done that =) and 3DS files. I
also load the 3DS hierarchy and animation, as you can see in the two
bottom-left pictures. In order to write this loader, I used some "old"
docs (from 1996 or 97) to be able to read the 3DS chunk file, and the
sources of Clax to
understand the behavior of rotation and pivot point in 3DS.
Link to the 3D face reconstruction page:
Link to my homepage: