Abstract
Modes of color experience refers to the phenomenological observation that two different color appearances can be simultaneously present at the same “location” as distinctive aspects of the percept (Mausfield, 2004). While these “modes” are often considered a product of higher order perceptual processes, here, we follow our previous studies (see Dixon and Shapiro, 2017) and explore the hypothesis that the spatial aspects of the stimulus contain much useful information not typically considered when building models of color vision. We argue that there are multiple spatial channels for color vision, and these channels are best suited for particular perceptual tasks - that is, the visual system separates any image into low spatial frequency (LSF) color channels, high spatial frequency (HSF) color channels, and rectified color contrast (RCC) channels. We demonstrate the advantages to this approach in three different ways: 1. We repeat demonstrations showing that the removal of low spatial frequency content can account for most brightness illusions (e.g. checker shadow, rubik's cube). We then show a computational demonstration that illustrates that even with very weak surround inhibition, a center-surround organization prevents the representation of LSF content. 2. We show that in many single illumination conditions, HSF channels are invariant to illumination. 3. We show that in many classic artistic representations, swapping low spatial frequency content is akin to swapping the apparent illumination. The results suggest that many contextual illusions and modes of appearance can be extracted from energy in the stimulus at different spatial scales.