Swapping swatches: Adapting to and from an artist's palette

We describe a method for representing and manipulating the color gamuts used by different artists to explore how the color schemes employed by artists might appear to the artist or to others. The method involves modeling the visual response to color and then adapting that response to simulate how color percepts change across different states of adaptation. Analyses of paintings and nature photographs suggest that there are both important differences and regularities in the color palettes of artists and that these regularities reproduce prominent characteristics of the natural color environment. In particular, the works of many artists include a bluish-yellowish bias that is also a distinguishing feature of both the color statistics of natural images and of the neural coding of color. The algorithm adjusts the colors in an image so that they are equivalent to the colors that would be experienced by an observer adapted to a different environment, or for two observers with different spectral sensitivities but who are adapted to the same environment. This provides a novel method for visualizing how the colors in artwork are experienced by an artist or an audience, and could be generalized to explore similar questions for visual attributes beyond color.