Word2Color converts natural language color descriptions to standard HTML4 colors. This is especially useful in data science, where color attributes often do not follow any normal list -- this enables users to quickly add various colors to simplifying bins of HTML4 standard colors. It does this by an input query (say "blue cream"), scraping the first X results from bing images (X is a parameter you can specify) for your query, and using K-Means Clustering to find the dominant color in each image, which is then averaged with the other images scraped and converted to the closest HTML4 color.

Word2Color is only tested on Python 2.7.X, although it likely works on Python 3 as well. OpenCV 2.4.X is a dependency, but is not listed in the requirements.txt to maintain compatability with virtual environments (See: How to install OpenCV in virtualenv). To install with pip, run:

pip install word2color

Note that after installng with pip, you should ensure you have all the requirements in the requirements.txt.

>>> from word2color import word2color
>>> word2color.color_description_to_bin('blue cream')
'silver'

Pull requests are welcome!

Word2Color is licensed under the MIT License.

Copyright (c) 2016-2016 Nelson Liu and Aakash Sethi