A comparison of the colorimetric characterization algorithms for different displays

Colorimetric characterization is the basic work of color reproduction. To find an optimized method to evaluate different types of displays is of great importance. The object of this study is to compare several universal algorithms which can be used for different types of displays. Different color space and whether gamma correction is applied are considered. The results illustrate both methods generated very consistent MSE for different displays. Polynomial modeling has a stable performance than neural network. Using 20 polynomial terms can get acceptable results for most displays. Gamma correction is not helpful to improve the accuracy, especially when neural network is used. Color space LAB is more stable than XYZ.