> When you try color metrics (measuring differences between colors),
> things get even more scary. Colorimetrisis found out that the human
> chromacity space IS 3D, but its metrics are non-euclidean....
> Do you still dare to scan? :-)
My company has built several color CCD-based cameras, and the customer's
are universally weird about the expectations:-) And I can't say I blame
them. Whereas a gray scale image of a gray document can be deemed "good"
or "bad" by engineering measurements, color vision is largely a psychological
phenomenon and the judgment of good vs bad is mostly aesthetic and/or
We found that when an observer is asked to judge the best from a set of
images, the "accurate" (in an engineering/colorimetry sense) image is almost
*never* chosen. An observer usually prefers excessively bright and slightly
saturated color images over crisp, precise images. And the scanners that you
buy (and televisions too, it seems) are optimized accordingly.
And from that perspective, the colors chosen in animated cartoons make
a lot more sense. (Purple hair?-)
And none of this is good news to someone who wants to create archival
images of precious documents. We've been there, too, and the conclusion
is that for all but the most extreme cases, good ol' 12bit JPEG (and an
environmentally controlled vault) are about as good as is worth the
effort. If some hot new image processing technique comes to light, you
are just going to have to accept the reality that you will need to open
the vault and rescan the document with the lighting/sensors required
of the technique.
-- Steve Williams "The woods are lovely, dark and deep. email@example.com But I have promises to keep, firstname.lastname@example.org and lines to code before I sleep, http://www.picturel.com And lines to code before I sleep."
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail email@example.com
This archive was generated by hypermail 2b29 : Wed Dec 13 2000 - 17:17:56 PST