Designing robust perceptual colormaps for Internet visualization

Abstract

Perceptual colormaps are colormaps that are designed using principles of human color perception. The importance of perceptual colormaps in visualization is that they provide a way to faithfully map data to colors without producing perceptual artifacts. Typically, a perceptual colormap is specified as path in a perceptual color space. To visualize the colormapped data on a display monitor, this path first needs to be converted to device RGB space of the specific display monitor. There are two problems that arise from the need to convert from device-independent to device-dependent color space. The first is that it limits the usefulness of perceptual colormaps to viewing environments in which we have control over the viewing factors that influence color perception. This excludes basically all Internet-based visualizations, since these are usually displayed on uncalibrated monitors under uncontrolled ambient lighting. The second problem is that the color space conversion methods are designed to achieve color fidelity, not perceptual fidelity, and therefore they are not particularly suited to creating perceptual colormaps. In this paper we consider the problem of designing perceptual colormaps that are suitable for the Internet-based visualization. We show that there are significant advantages in designing colormaps directly in device RGB space, rather than in a perceptual color space. In particular, this approach allows us to easily design "perceptually robust" colormaps, which maintain their perceptual properties over a wide range of viewing environments.