Re: diffusion limited aggregation

[Wayne Rasband]

> Hello.  I put this together for someone who wanted to try it in
> ImageJ, and the result is a working plugin that makes some
> limited simulations of diffusion limited aggregation.

There is an improved version of this plugin, plus four beautiful
fractal growth models generated by it, at
"http://rsb.info.nih.gov/ij/plugins/DLA.html".

-wayne

[audrey karperien]

It's much better now!  I really like watching it grow.  Audrey

Wayne Rasband <[hidden email]> wrote:
> Hello.  I put this together for someone who wanted to try it in
> ImageJ, and the result is a working plugin that makes some
> limited simulations of diffusion limited aggregation.

There is an improved version of this plugin, plus four beautiful
fractal growth models generated by it, at
"http://rsb.info.nih.gov/ij/plugins/DLA.html".

-wayne

[Wayne Rasband]

> It's much better now!  I really like watching it grow.  Audrey

It does that by having the BufferedImage and an ImageProcessor share
the same pixel data. I found that this sharing slows the plugin down
somewhat on Mac OS X.

Have you seen the DLA applet at
"http://apricot.polyu.edu.hk/~lam/dla/dla.html"?

There is a newer version of the plugin on the ImageJ website that does
a better job of estimating the needed image size. It use this formula

             neededSize = (int)(180+0.047* NumberOfIterations-4.57e-7*
NumberOfIterations*NumberOfIterations)

which I created using by running the plugin 6 or 7 times, measuring the
size of the generated model, and then using the Analyze>Tools>Curve
Fitting tool in ImageJ to fit a polygon. I found, the hard way, that
the plugin sometimes stalls if the image is too small.

-wayne