Controlling the Gray Component with Pareto-Optimal Color Space Transformations

An important and difficult problem in the printing and graphic arts industries is that of color management. Color management refers to the facilitation of color reproduction among various digital color imaging devices such as scanners, displays, and printers. The current work centers on color management for four color printing, focusing specifically on the practice of gray component replacement (GCR). Various techniques have been proposed for conversion into printer color spaces.1 Transformation into a printer color space typically requires conversion from a device-independent color space into the CMYK color space, where CMYK refer to the dot fractions of cyan, magenta, yellow, and black.2 Difficulty stems from nonlinear ink mixing behavior, gamut mismatch, and the indeterminacy of a 3-space to 4-space conversion. Techniques for performing transformation into the CMYK color space include: interpolation methods, models based on optics and ink mixing, and regression models. Interpolation methods involve the creation of a look-up-table (LUT) from which output values can be calculated using a variety of interpolation algorithms.3,4 Models based on optics and ink mixing include the Neugebauer equations, the Yule–Neilsen model, the Clapper–Yule model, the Kubelka–Munk theory, and the Beer–Bouguer law.5 Regression models typically involve finding model parameters (such as polynomial coefficients) which minimize the difference between a numerical model and a set of characterization data.6–8 Artificial neural networks (ANNs), which were utilized in this study, are one type of regression model that has been applied to color printing.9–13 The black printer is used in addition to cyan, magenta, and yellow due to a number of benefits. These benefits include: an increase in maximum obtainable density, improved stability throughout a press run, a reduction in cost as less expensive black ink is substituted for chromatic inks, and improved ink drying.14,15 In terms of transformation from a device-independent color space into the CMYK color space, the addition of the black printer has two main effects: extension of the printer gamut, and the introduction of redundant solutions. Figure 1 was created to illustrate the difference between the CMY and CMYK gamuts. Figure 1 contains both experimental data and model data; physically measured data are shown as circles, and data generated using the program NeuralColor are shown as solid points. The results show the gamut expansion that results from addition of a black printer and correlate well with the results of Nakamura and Sayanagi.16 There exist three regions in the CIELAB color space relevant to GCR in four color printing: the out-of-gamut region, the CMY gamut, and the region outside the CMY gamut but within the CMYK gamut. Colorimetrically accurate transformation from the out-of-gamut region is impossible and requires gamut mapping.17,18 The conversion of out-of-gamut colors to the closest in-gamut Controlling the Gray Component with Pareto-Optimal Color Space Transformations