A literature search showed Nayatani 's color appearance model to be inferior to Hunt's model. The goal of this research was to determine what modifications could be made to improve Nayatani's model. The coefficients' (xi), (eta), and (zeta), in Nayatani's model represent the chromaticity coordinates of the illuminant in the fundamental primary system. These coefficients were used in calculating the effective adapting level for each cone and act as normalizing factors i n the calculation of Nayatani color space. The CSAJ data set was used to determine a baseline for Nayatani's model, and results from modifications were compared relative to Hunt. Nayatani's Ml model set the coefficients' equal to 1, and Illuminant A was used to normalize the estimates of the fundamental primaries. Nayatani's results were compared to the CSAJ D65 reference data. The average AE*ab error was essentially equal for Munsell Value 3/, 5/, and 7/, while the total AE*ab error was reduced from 7.19 to 6.3. In addition, light colors shifted yellow and dark colors shifted blue for all three Value levels. This correctly predicts the Helson-Judd effect, and the results appear to match the Hunt model. The exponents of Nayatani's nonlinear p functions were rounded to the hundredths place to test the relevance of four significant figures. The (delta)L*, (delta)a*, (delta)b* and (delta)E*ab errors were the identical to those in Nayatani's standard model for all Value levels. Thus there is no justification for specifying exponents to four decimal places. Variations in the adapting luminance showed (delta)E*ab errors to be nonlinear at low luminance levels.
Library of Congress Subject Headings
Color vision--Mathematical models; Color vision--Computer simulation; Colorimetry
Department, Program, or Center
School of Chemistry and Materials Science (COS)
Rahill, John, "Computational analysis and revision of the Nayatani et al. appearence model" (1996). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus