Chromatic border perception: The role of red- and green-sensitive cones

Abstract

Where the perception of borders is concerned, the normal visual system is found to be tritanopic. The strength of a border depends on the relative activity of only R and G cones, no matter what the B cones are doing. Since normal color vision depends on the activity of all three cone types, chromatic border strength alone cannot be used as an index of color differences. Borders produced by spatial luminance steps or chromatic steps appear similar, except that Mach-type lateral inhibitory mechanisms, which enhance the visibility of low-contrast luminance borders, are unnecessary to predict chromatic border strength. Two functions are presented, using only the responses of R and G cones, which provide a way of expressing chromatic border strength in terms of either a subjective rating scale or a luminance step that appears to be equivalent in its distinctness.