Purchase this article with an account.
T. Kuyk, J. Dykes, M. LaFrance, L. McLin; Predicting Errors in Color Naming Caused by Laser Eye Protection. Invest. Ophthalmol. Vis. Sci. 2010;51(13):6299.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Laser eye protection (LEP) often acts like a colored filter by selectively absorbing visible light. Like colored filters, LEP can alter the appearance of colored stimuli. For aircrew, this can result in compatibility issues for in and out-of-cockpit viewing. This study investigated the effects on color naming of filtering at different locations in the visible spectrum and how well those effects were predicted by different color metrics.
Color naming data acquired over 10 years of testing with different LEP were analyzed. All observers who particpated had normal color vision. Color naming (CN) was assessed with and without subjects wearing LEP using a test where subjects identified the hue of 8 colored stimuli presented on a monitor. Fifty-four LEP that varied in the width and spectral location of the filtered wavelength band(s) and optical density were tested. The filters were designed to protect against lasers operating in the blue-green, green, yellow and/or red regions of the visible spectrum. Color difference metrics E94, Eab, Euclidian distance in uv space, and a metric based on cross-overs of color zone boundaries were evaluated for their ability to predict CN errors. For the zone analysis, color space was divided in to 8 regions (one achromatic). Chromatic zones boundaries were defined by hue angle and the achromatic zone boundary (from Kelly and Judd) by saturation relative to a D65 white point.
Baseline (no LEP) CN performance was near 100% correct for all 8 hues in the color set. With LEP, CN errors for some hues increased to nearly 100% misidentifications. Which colors were mis-named varied depending on the location of the absorption band(s). On a filter basis, the variables hue angle and saturation change in the color zone analysis, when combined, were significantly associated with CN errors (r2=0.78). Euclidian distance in uv space yielded an r2=0.75. Eab and E94 were less successful at predicting CN (r2= 0.62 & 0.69, respectively).
A metric based on color zone boundary cross-overs that combined information on hue angle and saturation changes was best able to account for color naming performance with LEP. A metric based just on hue differences was nearly as good. Both metrics need to be evaluated with broad band stimuli along with optimization of zone boundaries.
This PDF is available to Subscribers Only