Newsgroups: sci.med.vision,sci.image.processing,sci.med.physics,sci.med.informatics,comp.graphics,comp.graphics.visualization
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!udel!news.sprintlink.net!pipex!sunic!trane.uninett.no!ugle.unit.no!news.uit.no!bjoerns
From: bjoerns@stud.cs.uit.no (Bjoern Stabell)
Subject: Re: Chromaticity coordinates of Ishihara colors
Sender: news@news.uit.no (News admin.)
Message-ID: <1995Jan28.131502.18387@news.uit.no>
Date: Sat, 28 Jan 1995 13:15:02 GMT
References: <1995Jan22.174551.28522@news.uit.no> <poynton-2501951235120001@woody05.inforamp.net>
Organization: University of Tromsoe
X-Newsreader: TIN [version 1.2 PL2]
Followup-To: poster
Lines: 46
Xref: glinda.oz.cs.cmu.edu sci.med.vision:1741 sci.image.processing:12221 sci.med.physics:3116 sci.med.informatics:892 comp.graphics:69675 comp.graphics.visualization:7082

Charles A. Poynton (poynton@inforamp.net) wrote:
] In article <1995Jan22.174551.28522@news.uit.no>, bjoerns@stud.cs.uit.no
] (Bjoern Stabell) wrote:

] > I'm trying to synthesise how color-defective observers see
] > colors, and I want to evaluate the algorithm against well-known
] > color-blindness tests.

] >  ...    Does anyone know the color coordinates (in any device independent
] >         color model/space, but preferrably one defined by CIE) of
] >         the colors used in the Ishihara tests, or for that matter,
] >         _any_ color-blindness tests?

] Ishihara depends on reproducing the spectra of the stimuli, not just the
] chromaticity coordinates. To solve this problem I recommend that you start
] with my colour FAQ, recently posted here, then consult R.W.G. Hunt's
] Measuring Color, Second Edition (Ellis-Horwood), where various forms of
] colour deficiency are described. 

I've read them both.  And, according to several sources (e.g.,
Wysecki and Stiles), the dichromats have "confusion lines" that
radiate from a "confusion point" in the CIE x,y chromaticity
diagram; colors that lie on these lines can be matched by only
changing the lightness/luminance.

I guess this is what the Ishihara test uses; e.g., the colors are
chosen so that the colors that should be confused for a given
type of dichromat (and thus anomalous trichromats of the same
'class'; protan, deutan, and tritan) have the same lightness and
lie on the same confusion line (for 'class').

Lightness is "easy" to get right, but the chromaticity
coordinates tend to be distorted by scanning and printing
(calibration problems); sometimes the colors even lie outside the
input/output gamut of the devices, and so an accurate measurement
and/or reproduction becomes impossible.

I therefore need the chromaticity coordinates of (some) of the
plates on the Ishihara test specified; I would just use them to
evaluate my algorithm.


Bye,
--
Bjrn Stabell <mailto::bjoerns@acm.org>
              <http://www.cs.uit.no/~bjoerns/>
