Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!uwm.edu!linac!att!att!ulysses!andante!alice!jj
From: jj@alice.att.com (jj, like it or not)
Newsgroups: comp.dsp
Subject: Re: Mean Value Filter
Keywords: image processing, filter
Message-ID: <20094@alice.att.com>
Date: 22 Mar 91 14:25:17 GMT
References: <1991Mar20.003425.14767@sunee.waterloo.edu> <1991Mar20.224801.7413@leland.Stanford.EDU>
Reply-To: jj@alice.UUCP (jj, like it or not)
Organization: NJ State Home for Bewildered Terminals
Lines: 24

In article <1991Mar20.224801.7413@leland.Stanford.EDU> rick@hanauma.stanford.edu (Richard Ottolini) writes:
>In article <1991Mar20.003425.14767@sunee.waterloo.edu> rchann@sunee.waterloo.edu (Robert Chann) writes:
>>I'm trying to implement a mean value filter.  That is, using an nxn
>>approximately two additions per pel are required for the filtering.

>Simple.
>Use a rolling window.  Add elements of the leading edges and subtract elements
>of the trailing edges.


Yes, indeed, and watch the mean value slowly diverge in some direction
or other until its uncorrelated with the actual mean value.  Now, you
might get away with this for a 512 x 512 image or so, because you won't have
too much roundoff error, but don't try it on anything much bigger.

Don't forget roundoff error!

If you do the filter recursively (and that's what you're doing), 
don't forget that it has no controllability.
-- 
       -------->From the pyrolagnic keyboard of jj@alice.att.com<--------
Copyright alice!jj 1991,  all rights reserved,  except transmission  by USENET and
like free facilities granted.  Said permission is granted only for complete copies
that include this notice.    Use on pay-for-read services specifically disallowed.