IEEE Consumer Electronics Magazine - October 2015 - 75

(a)

(b)
FIGURE 3. A comparison between (a) an LDR image and (b) an
HDR tone-mapped image. [Images courtesy of Zicong (Jack) Mai.]

Photoreceptors' Response

ranging from 10-6 to 108 cd/m 2, we cannot perceive this range
simultaneously. Instead, our visual system needs to adapt to
the environment.
For our vision system to adapt between night and daytime,
it relies on two types of photoreceptor cells: cones (5 million)
and rods (125 million) [1]. Cones are responsible for the perception of color, finer details, and rapid changes. Three kinds
of cones with different peak wavelength sensitivities exist:
long-564 nm (red), medium-533 nm (green), and short-
437 nm (blue) [3]. The response of these cone types at different light wavelengths is illustrated in Figure 5. The weighted
combination of these responses allows our brain to differentiate colors. Note that cones can only operate in daylight and
stop functioning when there is not enough light entering the
eye. This is known as the photopic vision regime, ranging from
0.01 to 106 cd/m 2 .
Rods, on the other hand, are highly sensitive to light (i.e., the
number of photons) and are responsible for low light levels
known as the scotopic vision regime, ranging from 10-6 to
10 cd/m 2 . As there is only one type of rod, which is sensitive
around the blue-green wavelength, we experience a lack of color
at night. This property, i.e., the change of the color appearance
when illumination changes, is known as the Purkinje effect [4].
Cones and rods can operate simultaneously in low but not
quite dark lighting conditions (i.e., when the luminance level
ranges between 0.01 and 10 cd/m 2) . This is called the mesopic vision regime. Figure 6 illustrates the luminance range
where cones and rods are receptive as well as the different
vision regimes.
In addition to the fact that cones and rods are active at different luminance levels, they also adapt their response to the
ambient environment (see the variations of v in Figure 4).
Several experiments have been conducted to verify if this
property of HVS follows Weber's law, which states that when
the vision system is adapted to a uniform background, the
perception of a variation of intensity is usually linearly proportional to the background intensity. These experiments are
based on increasing the intensity of a gray patch on a uniform
gray background [5], [6] until the patch becomes visible and
a just-noticeable difference is reached. The experiments
shaped the threshold versus intensity (TVI) functions, which
proved that Weber's law is valid only when photoreceptors
are fully responsive (see Figure 7).
Another property of the HVS is its sensitivity to spatial
frequency at a low contrast intensity. The same contrast is
seen differently by our eyes at different spatial frequencies.
This effect is illustrated in Figure 8, where the contrast is
kept constant horizontally and the frequency of gratings is
increasing [8]. If the spatial frequency had no influence on
our perception of contrast, then the bar should vanish along a
unique horizontal line instead of a curve (see the dotted curve
in Figure 8). The highest pick the frequency at which the
human eye can distinguish the smallest contrast grating. Several experiments were conducted to measure the contrast
detection threshold for different spatial frequencies and background luminance [6], [9]. Those thresholds are known as

1
0.8
0.6
0.4
0.2
0
10−4

10−2
100
102
Intensity I in Log-Domain

104

σ = 0.01 and n = 1
σ = 1 and n = 0.5
σ = 100 and n = 1
FIGURE 4. The photoreceptors' response at different light intensities.

contrast sensitivity functions (CSF). Figure 9 illustrates
Barten's model of CSF [6] for several adapting luminance
values (L a) .

DIGITAL REPRESENTATION OF LUMINANCE
AND COLOR PERCEIVED BY THE HVS
For color perception, the HVS relies on three different types
of cones. To represent a reference color perceived by the
HVS, a weighted combination of three primaries is matched
october 2015

^

IEEE Consumer Electronics Magazine

75
Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - October 2015
