Each transformation has a weighted probability factor associated with
it. At each iteration the probabilities for each transformation are
evaluated and either applied or not, depending on the statistical probability.
After a reasonably large number of iterative cycles a new image will be
built up.
Here are some parameters to make the following fern image taken from
the fractint
ifs system:
fern {
| 0 | 0 | 0 | .16 | 0 | 0 | .01 |
| .85 | .04 | -.04 | .85 | 0 | 1.6 | .85 |
| .2 | -.26 | .23 | .22 | 0 | 1.6 | .07 |
| -.15 | .20 | .26 | .24 | 0 | .44 | .07 |
IFS fractals are very versitile and have been used to describe many
fractal shapes but also to describe many non-fractal images. Several different
image compression techniques were developed utilizing the "collage theorum",
put forth by Micheal Barnsley.
For other IFS sites, IFS utilities, and further information on this topic see:
For more information on Fractal Compression visit Yuval Fisher's
site on Fractal Image Encoding
at the Institute for Nonlinear Science University for California,
San Diego or the University of Waterloo, Fractal
Compression Project website.