The original Science News research report from 1999 is at:

The paper by a physicist, R.P. Taylor (which had to do with 
"authenticating" canvases by finding a characteristic (and fractal) 
pattern of, in this case Pollock's, work was printed in Nature 
(1999); an abstract may be found at: 
Part of finding this characteristic is distinguishing patterns from 
or detecting patterns in "noise"/randomness.  Similar methods 
(although I don't believe typically based on fractal geometries) are 
also used in determining authorship of written passsages.

There's a follow-up article "Fractal or Fake" from Science News in 
2007, which discusses some of the criticisms of the research (namely, 
that the fractal analysis doesn't prove what the original author 
claims it does) at:

Abstracts of the criticisms and responses, from Nature (2006) and 
Pattern Recognition Letters (2007) may be found at:

A preprint of Taylor's most recent statement and response from 
Pattern Recognition Letters may be found at:


At 9:46 AM -0500 6/8/07, Carrol Cox wrote:
>Diana Manister wrote:
>>  Peter, if you are referring to Jackson Pollock, he did not throw
>>  paint, he drew with it. His control was less than for an artist
>>  brushing paint on canvas, but that was deliberate. He wanted gravity
>>  and centrifugal force and other forces of nature to co-create the
>>  works with him. ("I am nature," he said.) Nevertheless, his own
>>  personal control was considerable. If you watch the films made of
>>  Pollock painting, you see that the applications of paint were never
>>  random, but carefully placed in relation to already existing shapes on
>>  the canvas.
>I unfortunately don't remember even the year let alone the issue of
>Science News which contained a story on how someone had done digital
>scans of Pollock's paintings and of actualy random 'paintings' and found
>very definite differences.