Abstract
Chromatic representations in the visual system beyond the cone-opponent level are still unclear. We applied the classification image (CI) method to examine the effects of heterochromatic noise on perception of chromatic contrast. The obtained data were analyzed by logistic regression on the basis of chromatic mechanism models, not just by the typical CI analysis procedure. The stimulus was a superposition image of two uniformly colored squares at left and right (a signal image) and a multicolored texture (a noise image) whose colors were randomly chosen. Observers judged the relative chromatic contrast of the two signal squares on different noise textures. The model analysis suggested that the cardinal-mechanisms model was enough to explain our experiment data just by adjusting the relative sensitivities between the mechanisms.
