Three-Dimensional Control of Color Management Systems - Control of Color Imaging Systems: Analysis and Design - page 306

7.3.4 ICC W ORKFLOW
The multidimensional pro

le LUTs embedded in ICC pro le have finite nodes at

which device values are available. Input image files with color pixels not on the

nodes will require interpolation. The ICC path provides simple linear interpolation
with limited resolution for processing pixels at high speed. When device behavior is
nonlinear as for the gamut of Figure 7.4, linear interpolation provided by ICC image
path may not provide suf

cient accuracy.

7.3.5 E NGINE C ONDITIONS
A stable engine condition is essential to repeatedly produce the color in the original
input file for the same media, toner, halftone, etc. Also, if the overall gamut of the

engine varies, then the colors initially matched to the original print are no longer
valid. As gamut boundaries move with time, static gamut mapping is no longer
optimal, leading to unpredictable results; even custom pro

les with static gamut

mapping will not reproduce optimal output quality.
The

le creation stage is device characterization. Nowadays,
characterization is performed by printing and measuring a set of color targets. The
next step involves computing the GCR-constrained inverse and gamut mapping
the out-of-gamut colors to the characterized boundary. All of these steps are too
demanding for high-end printing since even a small numerical error can lead to a
noticeable change in output color and a degradation in the image quality. Therefore,
in the following section we

first step in the pro

first describe the underlying theory and algorithmic
details of various methods applicable to characterizing a four-color print engine.
We show the adaptation process using recursive least-squares (RLS) identi cation

algorithms.

Nonlinear
color
surfaces

FIGURE 7.4 Nonlinear color surfaces (CYK L*a*b* gamut with color surfaces for
CY planes for fixed K and M ¼ 0).