Character Sequence Models for ColorfulWords

TL;DR

This paper introduces a neural network model that predicts color values from color names, outperforming human-created names in a preference-based test, with applications demonstrated through an online system.

Contribution

A novel neural network architecture for predicting color in color space from character sequences of color names, trained on large-scale online data.

Findings

Model's predicted colors are preferred over human-created names in a color Turing test.

Large-scale dataset enables effective training and evaluation.

Demo system available online at colorlab.us.

Abstract

We present a neural network architecture to predict a point in color space from the sequence of characters in the color's name. Using large scale color--name pairs obtained from an online color design forum, we evaluate our model on a "color Turing test" and find that, given a name, the colors predicted by our model are preferred by annotators to color names created by humans. Our datasets and demo system are available online at colorlab.us.