CSC5280
Project 1: Image Morphing
Dept.
of Information Engineering,
The
Chinese University of Hong Kong
This report is best viewed
in Internet Exploror.
Image Morphing / Photo Morphing, derived from
"image metamorphosis" [1], refers to the technique of producing a
sequence to show how image A can be naturally transformed to image B.
This technique has a lot of applications, such as film industry, which
leads to the explosive growth of the research [3].
Many image morphing algorithms can
be understood as combination of generalized image warping with a
cross-dissolve between two images. An example is shown in Fig. 1, where
Image C(t) is the morphed version at the
time t of Image A and B. The cross-dissolve technique used in this
project is a simple linear combination: C(t) = (1 - t) A(t) + t B(t), .
A
B
(1-t)
Warped A(t)
+
t
Warped B(t)
=
C(t)
Fig. 1
An example of image morphing
In this project, I implement
two algorithms: Beier-Neely algorithm [1] and S.Y. Lee's deformable
surface algorithm [2]. As I use the same cross-dissolve method
described as the above, this report will only focus on the warp part of
the two algorithms. The morphing results and animations are given in
Section 3. In addition, the improvement and the user guide of the user
interface will be given in the appendix.
