Page 91 - Computational Colour Science Using MATLAB

78 COMPUTING COLOUR DIFFERENCE

Lweight = 1 + (0.015*(meanL-50)^2)/sqrt(20 + (meanL- . . .
50)^2);
Cweight = 1 + 0.045*meanC;
Hweight = 1 + 0.015*meanC*T;

dl = dl/(Lweight*paral);
dc = dc/(Cweight*parac);
dh = dh/(Hweight*parah);

de = sqrt(dl^2 + dc^2 + dh^2 + rt*dc*dh);

Users may wish to modify the scripts or to convert them into other
programming languages. In order to facilitate testing of any implementations
of these colour-difference equations Table 5.1 has been provided which lists 10
pairs of samples that Luo et al. (2001) have designed for testing the CIEDE2000
equation. The tristimulus values in Table 5.1 are for the 1964 observer and
illuminant D65 (X ¼ 94.811, Y ¼ 100.000, Z ¼ 107.304). Table 5.2 lists the
n
n
n
colour-difference values for the CIELAB, CMC(1,1), CIE94 and CIEDE2000
equations. Table 5.3 provides more detailed information on the intermediate
stages for the CIEDE2000 equation.

Table 5.1 Data for testing implementations of colour-difference equations reproduced from
Luo (2001). CIE tristmulus values (illuminant D65 and 1964 observer) are provided for 10
pairs of samples. The standard and trial data are denoted by subscripts S and T, respectively
Pair X S Y S Z S X T Y T Z T

1 19.410000 28.410000 11.576600 19.552500 28.640000 10.579100
2 22.480000 31.600000 38.480000 22.583300 31.370000 36.790100
3 28.995000 29.580000 35.750000 28.770400 29.740000 35.604500
4 4.140000 8.540000 8.030000 4.412900 8.510000 8.645300
5 4.960000 3.720000 19.590000 4.665100 3.810000 17.784800
6 15.600000 9.250000 5.020000 15.914800 9.150000 4.387200
7 73.000000 78.050000 81.800000 73.935100 78.820000 84.515600
8 73.995000 78.320000 85.306000 69.176200 73.400000 79.713000
9 0.704000 0.750000 0.972000 0.613873 0.650000 0.851025
10 0.220000 0.230000 0.325000 0.093262 0.100000 0.145292