Read on Twitter
i have an idea
so we have an image, right? it's split into a red channel, a green channel, and a blue channel.
like so.
of course to get these channels i zeroed out the other two with multiplication. (R,G,B) * (1,0,0) = (R,0,0)
here's the red channel plus the green channel. as you can see, there's red and green, but no blue
it's NOT a yellow channel, since you can see there are red parts and green parts. it's just mostly yellow, since there's no blue.
now im not sure if you can get a yellow channel, but if you did, it'd be of the form (X,X,0), right?
where X and X are the same since yellow is (1,1,0), but with other colors it might be different
Anyways, we have green plus red. All colors are of the form (R,G,0)
Cyan is (0,1,1). If we hue shifted the green to cyan, we'd have (R,G,G), and now we technically have a full color image
Like so.
Or if we take red and shift it to purple, we get this
This is pretty much what "color blindness simulators" do I think
(R,G,G) is tritanopia
(R,G,G) is tritanopia
(R,G,R) isn't a kind of color blindness you can get. R and B aren't actually next to each other, so B will just get filled in by G
Here are (G,G,B) and (R,R,B). These two are the classic red-green colorblindness. There's actually two kinds (protanopia and deuteronopia), but they both look the same because R and G cells both look the same
*have almost the same wavelength
this is (R,B,B). It has almost the same hue as (R,G,G), since the linked channels are the same
here's (B,G,B). I like this one
