Based on the paper "CIECAM02 and Its Recent Developments" by color researchers Ming Ronnier Luo and Changjun Li, published in the book "Advanced Color Image Processing and Analysis" on the 26th May 2012. This MATLAB implementation was inspired by the python module "colorspacious" written by Nathaniel J. Smith, available here:
https://github.com/njsmith/colorspacious
Smith's very comprehensive python module "colorspacious" was used to generate the new perceptually uniform colormaps which are now default for MatPlotLib 2.0. These colormaps are also available for MATLAB here:
My goal was to provide functionality as simple as the commonly used Lab colorspace conversions, whilst providing a much more perceptually uniform colorspace. Note that I replaced calculations with a matrix inverse, e.g. inv(A)*b, with the recommended and numerically more precise A\b. See:
https://www.mathworks.com/help/matlab/ref/inv.html
Many users will want to convert from sRGB1 to J'a'b' uniform colorspace, just as easily as they might use RGB2LAB and LAB2RGB. I made this just as easy with the convenience functions srgb_2_Jab and Jab_2_srgb:
Jab = srgb_2_Jab(rgb)
rgb = Jab_2_srgb(Jab)
These use default values that are appropriate for sRGB (D65 illuminant, etc). Note that the sRGB inputs are MATLAB standard values between zero and one.
While most people will likely want RGB to Jab and back again, the main functions provide the following conversions:
XYZ_2_ciecam02 % XYZ100 to CIECAM02.
ciecam02_2_XYZ % CIECAM02 to XYZ100.
JMh_2_Jab % CIECAM02 subset to Jab.
Jab_2_JMh % Jab to CIECAM02 subset.
And for convenience in MATLAB (note that XYZ is scaled so Ymax==1):
srgb_2_xyz % sRGB1 to XYZ1.
xyz_2_srgb % XYZ1 to sRGB1.
Please pay careful attention to the function input ranges!
Of course there is no point in writing a conversion this complex without testing it thoroughly. The test values are those referenced in "colorspacious". The script is included so you can check it yourself.
Yes, I am well aware that the function names do not follow the common MATLAB practice: case sensitivity is required because the two characters "h" and "H" have distinct meanings, and the underscore helps to distinguish the digits in the case "ciecam02_2_XYZ".