I wrote something similar in javascript a couple years ago and have some suggestions:

Have a look at the WCAG contrast ratio and luminance functions to gauge similarities/differences between colors. If you are looking to extract a theme those formulas are gold!

What sort of hashing/tolerance are you using for grouping? Using the above you can specify a tolerance level, and use sqrt((r1-r2)^2 + (g1-g2)^2 + (b1-b2)^2) to get a diff and then compare that with the tolerance.

I helped write a library called colorific to detect a colour palette given an image. You might find it interesting. https://github.com/99designs/colorific

We blogged a bit about how it works too: http://99designs.com/tech-blog/blog/2012/05/11/color-analysi...

Hello, thank you for your sharing. It's awesome library and I think to use colorific in my project. we don't need to reinvent the wheel.

But it seems to need improvement to support Pillow, latest version of Python2, etc. I still didn't try but I'll examine it. Thanks.

Yeah, we built it to work well on logos and vector artwork, but it might not work so well for photographs. Anyway, see what you think. :)

I started to use colorific for palette. it's good, and i think to support also your project. thanks again.

I would have understood the title if it had just said written in Python. Most of the time I don't really care what libraries are used.

sorry, it's my carelessness. I fixed the description of github repository.

What is it?

Seems like you post a picture to it and it returns the colour palette used. Seems pretty useful if you're into that kind of thing.

Yes you're right. I wrote it to create color palette for company logo. It's useful for designers or theme creators.