Does dichromatic color simulation predict color identification error rates?

Purpose: Several algorithms are available to transform colored digital images into simulated dichromatic color perception. These algorithms can be very illustrative of the problems dichromats experience in discriminating colors. The purpose of this study was to determine whether one type of transformation could provide a quantitative account of error rates in identifying colors displayed on a computer monitor.

Methods: The experimental task required observers to identify the color of small rectangles displayed on a computer monitor within a black background. There were eight colors. The number of errors for each color was recorded. Four deuteranopes and five protanopes participated. Color differences were calculated using normal trichromatic and dichromatic values. The dichromatic color differences were calculated using the procedure developed by Brettel et al. [J Opt Soc Am (A) 1997;14:2647-55].

Results: The relationship between error rates and color differences calculated in either color space was fit by an exponential decay function. However, the fit provided by the dichromatic color differences was no better than that using color differences calculated in trichromatic color space and both regressions could only account for approximately 30% of the variance in the data.

Conclusions: Correlations between the error rates in identifying colors for dichromats and color differences were low-to-moderate whether the color differences were based on normal trichromatic color space or a dichromatic transformation. This finding suggests that it may be sufficient to calculate the color difference only in color-normal space to determine whether the colors will be confused by a person with a congenital color vision defect. Although computer algorithms are useful in illustrating color discrimination problems experienced by dichromats, they may not offer any advantage over typical trichromatic color spaces in predicting performance in color identification. The lack of any advantage may be due to how dichromats use brightness information to identify colors.