Page 175 - Computational Colour Science Using MATLAB

162 CHARACTERIZATION OF PRINTERS

testmat(i,13) = testrgb(i,2)*testrgb(i,2)*...
testrgb(i,3);
testmat(i,14) = testrgb(i,3)*testrgb(i,3)*...
testrgb(i,1);
testmat(i,15) = testrgb(i,3)*testrgb(i,3)*...
testrgb(i,2);
testmat(i,16) = testrgb(i,1)*testrgb(i,1)*...
testrgb(i,1);
testmat(i,17) = testrgb(i,2)*testrgb(i,2)*...
testrgb(i,2);
testmat(i,18) = testrgb(i,3)*testrgb(i,3)*...
testrgb(i,3);
testmat(i,19) = testrgb(i,1)*
testrgb(i,2)*testrgb(i,3);
testmat(i,20) = 1;
end

% implement the model for the test set
ptesttarget = testmat*a;

% convert the densities to XYZ values
for i = 1:144
ptesttarget(i,:) = exp(ptesttarget(i,:)).*white;
end

% compute the CIELAB Delta E values
de1 = zeros(144,1);

for i = 1:144
lab1 =; xyz2lab(testxyz(i,:),’d65___31’);
lab2 = xyz2lab(ptesttarget(i,:),’d65___31’);
de1(i,:) = cielabde(lab1, lab2);
end

% display the test results
perf = [min(de1) mean(de1) max(de1)]

The results of the third-order model are shown in Table 9.1, and it can be seen
that the performance is rather similar to the performance of the neural network.