Colori: Scaling the Measurement of Individual Perceptual Colour Geometry

The geometry of human colour perception has been studied for over a century, yet a full empirically grounded metric tensor for three-dimensional colour space has only recently become available. This creates a new problem: a single population-average metric cannot capture the reliable differences between individuals, nor the changes produced by adaptation state, context, or psychophysical paradigm, and estimating such variation requires far more data than conventional laboratory procedures can supply. We introduce Colori, a gamified colour-vision task that collects large-scale discrimination measurements on smartphones and tablets. Initial data show that Colori recovers major features of the recently established metric tensor. We outline how such data can be combined in hierarchical geometric models---a route from a single average metric toward a family of observer- and context-dependent perceptual geometries.