This project sought to discover the implicit colors of words, phrases, and concepts. A set of words was selected and run through a google image search. The color distribution of the first hundred images was analyzed and re-synthesized using a self-organizing map (SOM), a machine learning algorithm for clustering and visualizing data. The SOM was implemented using ofxSelfOrganizingMap.
After the google images were downloaded, a Gaussian mixture model was fit to the distribution of pixel colors. A self-organizing map of the color distribution was produced by randomly sampling from the resulting smoothed GMM.
Below is a series of results in categories: seasons, biomes, holidays, food, sky, film, politics, animals, and sports.