Hi Jolyon and Cedric,
I am failed to generate RNL colour maps by using dichromatic system (domestic dog, LW & SW). I used the pictures finished cone capture, acuity modelled, and RNL ranked filter. However, I tried trichromatic animals (human) and it works.
I’ve search for the colour maps on user guide, yet I found no information dealing with dichromatic animals.
Given that the axis of colour map in QCPA framework is the ratio of channels (i.e., mw:lw), I have few questions:
- Do the number of channels matter to create colour map, for example, at least three channels?
- Is it possible to create colour map with each axis representing one channel?
- Is it possible to extract the raw data (X and Y value)?
Many thanks!
Heya,
Indeed, to create a 2D or 3D representation of colour contrast in the RNL model (and thus, the colour map function), you need at least two opponent processes, i.e. 3 or more channels. In fact, the tool is rather specifically designed for tri- and tetrachromatic visual systems.
The tool is designed to have each axis correspond to an opponent channel, according to the RNL model. If you want to plot each channel as an axis, you will find a Maxwell triangle to be what you are after. This is indeed very popular in the literature and would be as easy as plotting the cone values in an n-dimensional plot using R, Matlab, etc.
The X, Y and Z coordinates of each cluster can be found in the QCPA output (e.g., ‘ROI Cluster results’ and others)
Cheers,
Cedric