Hi Everyone,

I have been a bit of a lurker for a while but I did make an intro a
while back but I did not see it pop up on the list. Anyway I'm
studying a Masters in Digtial and Photographic Imaging, my current
masters project is the Colour characterisation of a digital camera,
which involves making a GUI in Matlab to make this process a lot easier.

Regarding profiling a digital camera, my program/GUI goes almost all
the way to producing a ICC profile (I am just waiting for the new
verision of the Image Processing Toolbox from Matlab that will allow
me to do this). I am not familiar with the CS Raw calibration method
that Malcom mentioned. I understand that the same principles or
processes used to profile a scanner can be applied to a camera, ie
polynomial regression. This produces a set of coefficents that allows
the transformation of the image from a device dependent colour space
(RGB) to a device indpendent colour space (CIE XYZ).
 Making custom profiles for cameras is fine if they are to be used as
scanners !! If you want to use the camera for photographing natural
scenes with changing ligting conditions the situation gets much more
complex because of metamerism. Custom profiles are best used when the
lighting conditions are constant just like in a scanner. Or at least
make a new profile each time you change the lighting in the studio.
What I am working on right now is designed to be used in fixed or
constant lighting conditions for the reasons mentioned above.

Malcom I hope this helps a little or maybe I am talking about
something way different from what you had in mind.  I'm still wading
through all the literature on this topic and it is not easy, give me
until late september. Then I can talk with a LITTLE bit more authority
on the topic.

Oh by the way I am using the MacBeth colour checker chart and the
Kodak E-60 (IT-8), in general the larger the number of colour patches
you use the better the result.

Cheers
James

Reply via email to