info on gamma / math feature in ImageJ

Posted by Stuart Anderson-3 on Aug 24, 2006; 8:12pm
URL: http://imagej.273.s1.nabble.com/Could-ImageJ-be-ported-to-a-Mobile-Phone-tp3701757p3701759.html

Does anyone know where I might find information on the mathematics of the gamma
correction tool in the ImageJ Math menu?

I've been using the Plot Profile tool to produce spectral plots from images
obtained by a USB camera that's part of a CD-based grating spectrometer.  The
raw signals from the camera are linear at low intensities (doubling the light
intensity produces a doubling of the signals), but quite non-linear at higher
intensities.  I understood that a gamma correction might fix this roll-off, and
it does help.  With my camera a gamma of 1.4 seems to be about right.  That is,
after a gamma correction of 1.4 the Plot Profiles of the corrected images do
double when the intensity is doubled.

However, the odd thing is that the gamma process, applied to the image before
the Plot Profile operation, yields results that are not one-to-one within the
Profile.  That is, a given value in the Plot Profile of the unprocessed image
can correspond to several different values in the Plot Profile of the gamma-
processed image (including zero and non-zero values).  I can see how - since the
gamma is applied to the image before the Plot Profile is calculated - this might
be the case, especially if the gamma correction is different for each of the RGB
components, but I'd like to understand this better than I do.

So, if anyone knows where I can find documentation on what the ImageJ / gamma
process does that would be very useful.

Thanks - Stu