I think this is a Coupled Map Lattice of the Logistic Map. Either way it's pretty. The logistic map is xn+1 = rxn(xn-1), and in this example, we're representing x as a shade from black=0 to white=1. The twist here is that on each iteration, instead of using the result straight away, it processes the results to add smoothing and then noise. The smoothing step: it takes the pixel itself and the pixels above, below, left and right of it (if they exist), and calculates an average. It then sets the pixel value to k times this, plus 1-k times the value before averaging. Finally, noise is added, by generating a number from 0 to 1, multiplying by noise parameter, and adding that to (1-noise) times the pixel value.
xx = x - 0.5; yy = y - 0.5; Math.random()*Math.sqrt((xx*xx)+(yy*yy))*1.4
4 - (y*y)
Peter Corbett, 2014
Back to ptc24 labs.