Light scattered in the eye and its effect on the measurement of the colour constancy index

Abstract

It is well established that at least some of the neural processes that mediate colour constancy effects are very rapid and practically instantaneous with the change of illuminant. This aspect of colour constancy has been labelled instantaneous or immediate in order to distinguish it from other factors, such as adaptation, which also contribute to colour constancy and have a longer time course. Instantaneous colour constancy has been studied using either simultaneous or dichoptic presentation of a test and a comparison field surrounded by a Mondrian pattern or a uniform background. In order to avoid local adaptation to each of the two illuminants, we developed a dynamic colour matching technique, based on sequential changes of illuminant which affects only the surrounding background and not the test target. This technique makes it easy to measure conveniently an index of constancy on a scale of zero to one, and is based on the change in the target chromaticity needed to keep the perceived colour of the test patch invariant under each of the two background illuminants. Using this dynamic technique, we were also able to investigate the spatial properties of immediate constancy effects using a dark annulus of variable width which separates the test target from the surrounding background. The results show that the effect of the immediate, adjacent surround is very important, although a significant C-index value can also be elicited with distant surrounds when the light scattered in the dioptrics of the eye from the surrounding background onto the test target is very small. The computations involved in such measurements are normally carried out on changes of chromaticity as measured on the visual display The forward scatter of light caused by the dioptrics of the eye can change significantly the chromaticity of the test target as imaged on the retina even when the test chromaticity measured on the screen remains unchanged. In this study we have investigated how scattered light in the eye affects the index of colour constancy. The scatter function which describes the amount and the angular distribution of light scattered in the eye has been measured for each subject. C-index values have been computed with and without correction for the changes in test target chromaticity caused by scattered light. The results show that scattered light in the eye cannot explain immediate constancy, and that on average its effect is to reduce significantly the estimated colour constancy index.