> -----Original Message-----
> From: [EMAIL PROTECTED] [SMTP:[EMAIL PROTECTED]]
> Sent: Tuesday, December 12, 2000 8:19 AM
> To: [EMAIL PROTECTED]
> Subject: Re: filmscanners: VueScan 6.3.19 Available
>
> In a message dated 12/11/2000 8:19:34 PM EST, [EMAIL PROTECTED]
> writes:
>
> > Now I'm confused (not hard to do). Please explain to me "where" ICE
> > is.....is it software or hardware? Or a combination of both to get
> that
> > capability?
>
> ICE is the marketing name for a dust removal technique invented
> by Albert D. Edgar while he was working at IBM (he's at Applied
> Science Fiction now).
>
> The patent can be read at:
>
> http://www.delphion.com/details?&pn10=US05266805
>
> It's U.S. Patent 5,266,805
>
> This patent only describes the way the infrared channel
> is used to correct the image, but it has several problems.
>
> The first problem is that it assumes the infrared channel
> doesn't show any image data, but in reality the infrared
> channel isn't flat (especially for Kodachrome). The
> second problem is that the infrared and color channels
> aren't perfectly aligned, which causes the edges of dust
> spots to not disappear when the algorithm in this patent
> is used.
>
> VueScan uses an entirely different (and I think better)
> approach to using the infrared channel to remove the
> dust spots. It doesn't result in any color shift, it works
> with Kodachrome, and it doesn't apply a softening filter
> in areas where there's no dust spot.
>
> Regards,
> Ed Hamrick
[Oostrom, Jerry] A ha, so Vuescan cleaning still uses the infrared
channel. Perhaps you remember from the thousands of mails received this year
that some person (I) once sent you a request for a cleaning algorithm that
probably does not exist yet. I thought this up myself, but perhaps a lot of
others did too and were just knowledegable enough to know it was asking for
the impossible or could never work.
Now I think I should propose it to this group and receive answers
from them why it is not feasible.
(Keep in mind that I am not good at math and such).
What I noticed with dust on slides in a neutral scan is that most of
the dust is real pitch black. With negatives it is black in the raw scan,
but has a color shift related to the white point chosen in the positive crop
file. Many times however, it will be at the boundary of the histogram.
Can an algorithm be constructed that applies selective
softening/cleaning at parts that are at the lower boundary of the histogram
for a slide and at the upper boundary of the histogram of a negative?
Has it already been made (does the clean filter in the newest
Vuescan already work as such)?
Did I ruin any surprise for a new cleaning algorithm in vuescan that
can be used for the wicked scanwit and any other scanner without ICE?
Please tell me if this idea is any good to investigate and build
upon,
thanks,
Jerry.
