Calculate how humans perceive similarity between different colours

Question

I\'m working on a site where users can describe a physical object using (amongst many other things) any color in the rgb 0-255 range. We offer some simplified palettes for e

Answer 1

There are a half dozen or so possibilities. EasyRGB has a page devoted to them. Of those listed, DeltaE 2000 probably has the best correlation with human perception -- and is also extremely complex to compute. Delta CMC is almost as good for something like half the code (though the computation still isn't entirely trivial).

Answer 2

This sounds like a prime example for a neural net based approach (if you are in an experimenting mode :) because it's about creating a decision rule that mimics Human perception. A neural net that has six inputs (r, r', g, g', b, b') and one output (is_similar) can be easily trained by using e.g. your own perception of similarity as the training source!

Answer 3

I'm not 100% clear on how your problem is set up, but you may want to read up on: Normalized Cross Correlation, and Lab and CIEXYZ color spaces.