END OF AN ERA, FRACTALFORUMS.COM IS CONTINUED ON FRACTALFORUMS.ORG

I found excelent plugin Fractalius. Look, this image before:



end this is after Fractalius:



I made other images in Fractal Explorer with Fractalius:





P.S. Sorry, how i can made normal image's size?

ahh, it's all becoming clear now... it's due to the misnaming of this plugin that (other) people think the images it produces are fractal! there are quite a few mis-categorised images like this on deviantart and elsewhere (a friend happened to show me one just seconds before i saw this thread, weird synchronicity); and one can understand the confusion - the naming for one thing, and aren't all shiny glowy swirly images fractals? </sarcasm>

the technique employed is called line integral convolution, and it sure does look good when applied to fractals (however it will tend to mask the details of self-similarity).

so, THAT's how they are created.... Always wondered about those glowing animals on dA lol...

I don't see how lic can be employed here.

I see "radial flares" in fft from fractalius. Wat it mean?

Your own images above should serve as examples, no?

Sure they are: you get the angle along which to perform LIC from the gradient of the luminance, using this you can construct normalised sampling vectors (using sin/cos/sqrt) and do the weighting along those "rays" using a Gaussian falloff function exp(-(k * dist)^2).

It's not absolutely essential to use sin/cos/sqrt/exp, but their use for LIC can easily be explained as above.

Have you implemented LIC before? I can give simple pseudocode, maybe 10-20 lines, that would produce near identical images to what's produced by Fractalius.

I'm at work right now, will post it tonight

sampled met exp, layered.

do sobel in sobel;
from sincos(sobel; +random(angel)) do line.blur.length.fn(hermite from intensity) in double accumulator; //re gauss rotate and ...
perform expmap in layers: