Abstract
Purpose: :
To estimate the effect of photopigment optical density on the ability of anomalous trichromats to discriminate colors in real-world scenes.
Methods: :
The cone sensitivities of our theoretical observers are modeled using the Lamb (1995) photopigment template and are corrected for the optical density (OD) of the photopigment and for the filtering effect of lens and macular pigment. In this way we can generate corneal sensitivities for cones containing photopigment of any peak wavelength, and at any optical density. In the model, we vary the "OD disparity", which we define as the difference in OD between the photopigments in the anomalous trichromat equivalent to L and M, and calculate as (OD of longer wavelength photopigment) - (OD of shorter wavelength photopigment). For seven real-world scenes and for a range of theoretical anomalous observers, we use the hyperspectral images of Foster et al. (2004) to estimate the signals that are produced in post-receptoral channels that are analogous to the L/(L+M) and S/(L+M) channels of the normal observer. The richness of color vision is estimated by determining the gamut of signals in a space defined by these two channels.
Results: :
Positive OD disparities increase the gamut of signals in our color space: For low peak separations an OD disparity of +0.1 provides the same benefit to color vision as would an additional 1 to 1.5nm of peak separation for both protanomals and deuteranomals. Small negative OD disparities are detrimental.
Conclusions: :
Positive OD disparities of up to 0.2 can be inferred from studies of dichromats, and similar disparities have been suggested for deuteranomals (Shevell and He, 1997). Our work suggests that such a disparity could explain why some anomalous trichromats enjoy substantially better color vision than we would expect from estimates of their peak separation.
Keywords: color vision • color pigments and opsins • discrimination