Xref: utzoo comp.dsp:1445 comp.graphics:16748
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!apple!netcom!mcmahan
From: mcmahan@netcom.COM (Dave Mc Mahan)
Newsgroups: comp.dsp,comp.graphics
Subject: Re: Mean Value Filter
Keywords: image processing, filter
Message-ID: <1991Mar21.005403.15906@netcom.COM>
Date: 21 Mar 91 00:54:03 GMT
References: <1991Mar20.003425.14767@sunee.waterloo.edu>
Organization: Dave McMahan @ NetCom Services
Lines: 34


 In a previous article, rchann@sunee.waterloo.edu (Robert Chann) writes:
>Hi,
>
>I'm trying to implement a mean value filter.  That is, using an nxn
>window, each picture element is substituted by the mean value of the 
>(nxn) neighbouring picture elements.
>
>Apparently, by using a "sophisticated" implementation, no more than
>approximately two additions per pel are required for the filtering.
>The number of addition is nearly independent of the filter window
>size and no multiplication is necessary.
>
>Does anyone know how this "sophisticated" implementation work?

Well, I think so.  I understand 'mean' to be the 'average'.  This means that
a pel will be the 'average' of all it's neighbors.  I know how to do it with
one addition per pel, plus overhead required to get at the numeric value for
each pel.  All you have to do is add them together, and divide by the number
of pels.  This will be independent of kernal size.  Most of the overhead
you will run into will be converting whatever graphics image format into a
regular old binary value that can be summed.  This can be easy or hard, but
depends entirely on the format of your data.

Is this what you want?

>Robert Chann
>(rchann@sunee.waterloo.edu)

  -dave

-- 
Dave McMahan                            mcmahan@netcom.com
					{apple,amdahl,claris}!netcom!mcmahan