Page 124 - Computational Colour Science Using MATLAB

7

Characterization of Computer

Displays

7.1 Introduction

Linear transforms are fundamental to the study of colorimetry and have many
applications, especially in the characterization of imaging devices such as
monitors, cameras and printers. Device calibration is concerned with setting the
imaging device to a known state and ensures that the device is producing
consistent results. Characterization is the relationship between device coordi-
nates (usually RGB or CMYK) and some device-independent colour space such
as CIE XYZ (Fairchild, 1998; Johnson, 2002). Green (2002b) argues that there
are three main methods for achieving this mapping: physical models, look-up
tables and numerical methods. Physical models often include terms for various
properties of the device such as the absorbance, scattering and reflectance of
colorants. The Kubelka–Munk model is an example of a physical model that can
be used as the basis of a characterization method for a printer (Kang, 1994;
Johnson, 1996). Similarly, the gain–offset–gamma model (also known as GOG)
is a physical model of a computer- or visual-display unit based on a cathode-ray
tube (CRT) that can be used for the characterization of most display monitors.
Look-up tables define the mapping between a device space and a CIE colour
space at a series of discrete measured coordinates within the colour space and
may interpolate the values for intermediate coordinates. For numerical methods
a series of coefficients is determined, usually based upon a set of measured
samples, without prior assumptions about the physical behaviour of the device or
its associated media. Examples of numerical methods include linear transforms,
non-linear transforms or polynomials and artificial neural networks. A key
property of any transform is whether it can easily be inverted. The advantage of
a linear transform is that it is trivial to invert whereas many empirical models are
not easily inverted (Iino and Berns, 1998). If inversion is not possible, then

Computational Colour Science Using MATLAB. By Stephen Westland and Caterina Ripamonti.
& 2004 John Wiley & Sons, Ltd: ISBN 0 470 84562 7