Page 18 - Computational Colour Science Using MATLAB

A BRIEF REVIEW OF THE CIE SYSTEM OF COLORIMETRY 5
vision is mediated by the responses of the cones, of which there are three types,
with sensitivities peaking at 420 nm (short wavelengths), 530 nm (medium
wavelengths) and 560 nm (long wavelengths), termed S, M and L cones,
respectively (Bowmaker, 2002). The three classes of cones are not distributed
evenly throughout the retina (Williams et al., 1981). In the central or foveal
region, for example, only L and M cones are present and there are approximately
twice as many L cones as there are M cones. The S cones are rare throughout the
retina but are more concentrated in a ring around the fovea. The retina contains
several layers of cells and the signals generated by the transduction of light into
chemical and electrical energy in the cones activate the bipolar and ganglion cells
before leaving the eye via the optic nerve. The cones and rods each provide a
univariant response (Wandell, 1995; Westland, 2002) and the consequence of this
is that the individual classes of cones and rods are colour blind. That is, the
scotopic visual system can only perceive shades of grey and the three classes of
cones considered separately are also incapable of wavelength discrimination. At
least two classes of cones are required for colour vision. The photopic visual
system achieves colour vision by analysing the relative responses of the three
classes of cones in the eye.
The CIE developed a system for the specification of colour stimuli that was
recommended for widespread use in 1931. The most important principle that
allowed this development was additive colour mixing. Thus, all colour stimuli
can be matched by the additive mixture of three appropriately chosen primaries.
It had long been recognized that the amounts or intensities of the primaries
required to match a given stimulus effectively form a specification of the stimulus
in terms of the primaries that are used. The amounts of the primaries used for
any given stimulus are commonly known as the tristimulus values. It is possible
to determine the tristimulus values for any given stimulus using a device known
as a split-field or bipartite colorimeter. In such a device an observer views a
bipartite field. On one side of the field the stimulus is displayed; on the other side
the additive mixture of the three primaries is displayed. The observer adjusts the
intensities of each of the three primaries until the additive mixture is
indistinguishable from the stimulus. Under the matching condition the field
appears uniform to the observer and the tristimulus values can be read off from
the device and recorded.
The measurement of the colour-matching functions by observers was a
critical feature in the development of the 1931 CIE system of colorimetry
since it allowed the computation of the tristimulus values for a known
stimulus without the need to view the stimulus in a bipartite colorimeter. The
colour-matching functions are the amounts of three primaries required to
match one unit of intensity of a single wavelength of light, and were recorded
for small wavelength intervals throughout the visible spectrum. If red, green
and blue primaries are used and these are denoted by the symbols [R], [G]
and [B], and the tristimulus values are represented by the symbols R, G