I wanted to measure the average color values of a small area of a photo, so I ran the QCPA framework and converted that cone-catch image to RNL chromaticity. I then measured that photo to get the average X and Y color values. Is it correct to say that these values are how “red” or how “yellow” the area is (x and y, respectively), and the farther away from 0 the value is, the greater the color intensity? If not, is there a way to easily transform these data to reflect color intensity? Thank you so much.
Please see here for how to interpret the X and Y coordinates of a tri-chromatic visual system in the RNL colour space. As far as the RNL model is concerned (and general definitions of colour) we can discern between hue and saturation. So I am not sure what ‘colour intensity’ would refer to? Usually, the term ‘intensity’ is used as a synonym for luminance which refers to how much light is reflecting from a surface which then is perceived as brightness and/or lightness (note: not the same and not of concern to the RNL model). Luminance and saturation are often intertwined, both in terms of how visual systems process these properties neurophysiologically, but also in our language. When we talk about a ‘bright red’ object, we often mean both, a very saturated red and a red with a high brightness/lightness. To get the parameters you are after, please pick the QCPA output parameters of choice, explained in detail here (e.g. VCA.MSsat) or, using the output of the ROI Cluster Result (e.g. RNL chromaticity, which is the distance to the achromatic point in delta-S). In short, yes, the X and Y axis coordinates can be used to describe how much a given opponency channel is stimulated. However, it is important to keep in mind that, just because i.e. the X-axis would indicate a strong stimulation of the ‘red’ photoreceptor, doesn’t mean the perception of a ‘red’ hue is the case. The other opponent channel needs to be incorporated into that analysis as well. i.e. a strong ‘red’ plus a strong ‘yellow’ response, in human visual systems equal an orange hue. Hence the name of ‘chromaticity’ for delta-S (it’s the combo of hue and saturation). I hope I was able to answer your question?
^ what Cedric said, but perhaps put more simply, assuming you’re using a trichromatic visual system (where you haven’t altered the arrangement of receptors – they should go from LW to SW) then:
X values code for a LW vs MW opponent channel (can be thought of as red-green, where positive values are red, negative are green – or vice-versa – I forget off the top of my head).
Y values code for LW & MW together vs SW channel (can be thought of as blue-yellow, as above, positive vs negative).
The central point is grey (achromatic point), based on the grey standard you used in your image.
Note that this is all independent of luminance or overall intensity, it’s all relative chromatic space.
These opponent channels are orthogonal (independent), which is mathematically convenient for us, but as Cedric says, visual systems will combine these channels. In practice, in terms of statistical analysis of colour, that means it could be inappropriate to ignore one of these channels, particularly the one showing the greatest variance with your experimental treatment (if one channel isn’t showing any variation you could justifiably drop it from your analysis though – you’d need to demonstrate this statistically though).