info

::

cv
Parameterization of 3D objects

2007 - Degree Thesis

The main purpose of this degree thesis was to implement and describe an innovative technique for the transformation of an image to another one and thus a technique that face recognition can be based on.
The idea was to parameterize a 3D object with natural models of deformation which incorporate both the shape (x, y) and the intensity I(x, y) components of a 2D image.
This method allows us to deal with the optical variation of the subject. The main idea is to deform the intensity surface of a source image into the intensity surface of a target image.
Specifically, the intensity surface is modeled as a deformable 3D mesh in (x, y, I(x, y)) shape which obeys Lagrangian dynamics. Using an efficient technique for matching two surfaces we can obtain a dense correspondence field (or 3D wrap) between two images.
This project was implemented using Matlab.