Abstract
We often perceive a color that is inconsistent with the light reflected from an object. For example, we can perceive constant colors despite changes in illumination. Therefore, it is considered that our visual system would correct the estimates of object colors with the statistics of surrounding colors. Recently, reverse correlation, or Classification Image analysis, has been used to reveal how the scene statistics affect color perception. However, there is a possibility that these conventional methods provide biased estimates when higher-order scene statistics affect color perception. Here we investigates the influences of surrounding colors on apparent color shift with a reverse correlation method considering the nonlinearity underlying observer's decisions.
