Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cica!gatech!bbn!inmet!rich
From: rich@inmet
Newsgroups: comp.graphics
Subject: Re: SigGraph Fractal Compression
Message-ID: <20400002@inmet>
Date: 16 Aug 89 15:53:00 GMT
References: <2037@netcom.UUCP>
Lines: 33
Nf-ID: #R:netcom.UUCP:-203700:inmet:20400002:000:1831
Nf-From: inmet!rich    Aug 16 11:53:00 1989


> Well, this is the best explanation that I've heard so far, but I'm
> still not convinced of the method's overall usefullness.  First, how
> does one go about finding those coefficients?  Is it an automatic
> process, or does it have to be done "by hand?"  How does this method

	Well, this is the "secret" that Barnsley is protecting.  It usually is
done in a brute force manner.  However, it used to take perhaps 100+ hours
to find the encoding, but he claims his new method takes perhaps minutes to
several hours.

	This is not a standard compression scheme.  The difference is instead of
compressing the bitmap, you are giving instruction of how to draw the image
using some standard shapes: in CAD areas, it may be circles and squares etc,
for "natural" images, you use fractals.  SO depending on how many fractal
equations you need, that's your compression ratio.

> As far as the getting more resolution out than you put in, surely you
> realize that this is nonsense.  You are certainly able to do the
> calculations, but the added information is just arbitrary and won't
> necessarily have anything to do with the original information (I don't
> think).  In order to reproduce the image at the original resolution,
> you only need a certain amount of "accuracy" for your coefficients.

	What I meant is you can reproduce your image at any resolution.  The 
fractals equations are resolution independent.  So this scheme works as an image
enlarger/reducer.

	I have my doubt whether we can really encode images in real time w/ this 
technique.  However, I think this technique is very useful for archiving images
and in situations that transmission of images is more expensive (such as say
from a space probe).  BTW, the system is more error "resistant".  That is, a 10%
error on data yeilds less than 10% error on image.