color compatibility from large datasets - university of.pdfVIP

Color Compatibility From Large Datasets Peter O’Donovan Aseem Agarwala Aaron Hertzmann University of Toronto Adobe Systems, Inc. University of Toronto Abstract Our studies are based on three datasets, each of which comprises a collection of color themes and their ratings. We derived two datasets This paper studies color compatibility theories using large datasets, from Kuler and COLOURLovers, and created the third using Ama- and develops new tools for choosing colors. There are three parts zon Mechanical Turk (“MTurk”). These datasets exhibit different to this work. First, using on-line datasets, we test new and existing advantages and disadvantages. For example, Kuler users have more theories of human color preferences. For example, we test whether exposure to color theory than MTurk workers, while MTurk data certain hues or hue templates may be preferred by viewers. Second, is collected in a more controlled fashion. However, taste in color we learn quantitative models that score the quality of a set of five can vary widely, and users in these datasets have varying goals, colors, called a color theme. Such models can be used to rate the backgrounds, and viewing environments; not surprisingly, there is quality of a new color theme. Third, we demonstrate simple proto- substantial variation. Nonetheless, analysis of the data reveals many types that apply a learned model to tasks in color design, including regularities and patterns. improving existing themes and extracting themes from images.