Energy efficient algorithms and techniques for wireless mobile clients 3

CHAPTER 2
RELATED WORK
Power management in battery operated mobile devices has been an active area of
research. There are several proposals that attempt to reduce the power consumption
of various components of the mobile devices. In this chapter, we review a range
of techniques available in the literature, primarily focusing on display, network and
processor components.
2.1 LCD Power Conservation
Most commonly, the Liquid-Crystal (LC) cells forms the pixels of Liquid-Crystal
Displays (LCDs). These cells can react to the modulation of electricity fields and
change the polarization direction of light passing through them in response to an
electrical voltage. Thin Film Transistor LCD (TFT LCD) has a sandwich-like struc-
ture with LC filled between two glass plates as shown in Figure 2.1. TFT glass has as
many TFTs as the number of pixels displayed, while a Colour Filter glass has colour
filter which generates colour. LCs move according to the difference in voltage between
the Colour Filter glass and the TFT glass. These LCs are aligned electronically to
form a pattern or image. Based on this pattern the backlight passed or blocked to an
outer layer (transmissive layer) to create the image.
18
Figure 2.1. Structure of a Transmissive TFT LCD
Most of the LCD power is consumed by its Backlight. There are a several works
to reduce the backlight power consumption. Dynamically dimming the backlight is
considered an effective method to save energy consumed by the mobile device dis-
plays. The resultant reduced brightness can be compensated by image enhancement
techniques such as, scaling up the pixel luminance. Traditional LCD displays used
Cold Cathode Fluorescent Lamps (CCFL) for the backlight while modern displays
use Light Emitting Diode (LED) arrays instead. Dynamic dimming techniques de-
scribed below can be applied for both CCFL and LED based backlights. However,
as LEDs consume less power than CCFL, LED based LCDs are power efficient than
CCFL based LCDs [44]. A 3-in-1 RGB LED can obtain a wider colour gamut and
better pre-mixed colours.

Various techniques in the literature for conserving LCD backlight energy are dis-
cussed below.
19
Figure 2.2. Visibility of the Image in a Transmissive TFT in some Environment
Luminance Condition [2]
Backlight Auto-regulation
In Gatti et al. [2], a technique called backlight auto-regulation was devised which
regulates the backlight based on ambient lighting levels. The relation between re-
quired backlight luminance and ambient lighting is shown in Figure 2.2. This tech-
nique employs an on-board environment luminance sensor to determine the appropri-
ate backlight needs based on an ambient light condition. The authors adjusted the
input voltage of the backlight driver to modify the luminance. Power saving ratio
becomes more significant, reaching up to 74% as the environment becomes darker
and the backlight luminance becomes lower.
This is a very simple backlight dimming technique and it is already implemented
in almost all modern smartphones with some better calibrations. However, this tech-
nique can be complemented with image enhancement technique to save more energy.
20
Figure 2.3. Image and its Discrete Histogram
DLS - Dynamic Backlight Luminance Scaling
The principle of DLS is to save power by backlight dimming while restoring the
brightness of the image by appropriate image compensation. This scheme is generally
known as Dynamic Backlight Luminance Scaling. In Chang et al. [33], a dynamic
backlight luminance scaling scheme is proposed. Based on different scenarios, three
compensation strategies are discussed, that is, brightness compensation, image en-
hancement, and context processing. Brightness compensation results in image dis-
tortion. Hence, amount of increase (transformation) to pixel luminance is controlled
by a threshold T
H
(shown in Figure 2.3) which is, determined using distortion ratio

(D
i
). Distortion ratio (D
i
) is given in Equation 2.1.
D
i
=

2
n
−1
j=T
H
H
j
(M
i
)

2
n
−1
j=0
H
j
(M
i
)
(2.1)

where, M
i
represents the image or frame as a matrix of pixels, H is histogram
function,H
j
represents histogram value of a particular colour j.
21
Brightness compensation allows a significant degree of backlight dimming while
keeping the distortion ratio reasonable, as long as the image has a continuous his-
togram (adjacent histogram values are close to each other) which is not severely
skewed to bright areas. Brightness compensation is not efficient for images with
discrete histogram. For discrete histograms Image enhancement techniques such as
histogram stretching and histogram equalization are proposed. The transformation
to the pixels are controlled by lower threshold T
L
and upper threshold T
H
as shown
in Figure 2.3. However, some minor colours may be merged into each other and are
thus no longer distinguishable after histogram equalization. This may result in small
parts of images which has similar colour as background to become indistinguishable
from each other. Hence, a context based processing is proposed.
However, their calculation of the distortion does not consider that the clipped
pixel values do not contribute equally to the quality distortion.
CBCS - Concurrent Brightness and Contrast Scaling
In Cheng et al. [34], a similar method, namely, Concurrent Brightness and Con-
trast Scaling (CBCS), is proposed. CBCS aims at conserving power by reducing the
backlight illumination while retaining the image fidelity through preservation of the
image contrast. The authors defined a contrast fidelity function (f
c

(x)) for measur-
ing the image fidelity after backlight scaling. The authors also added a logic which
controls the distribution of output voltages to an original voltage divider in imple-
menting a scaling function of LCD’s transmissivity. This consequently eliminated the
pixel-by-pixel manipulation on the image. A voltage divider is a specific hardware
22
designed to produce fixed voltages required by a source driver for setting a certain
level of LCD’s transmissivity.
Cheng et al. [34] clearly modelled the observed luminance of a transmissive
object as a product of the backlight luminance and transmissivity of TFT-LCD. The
modified voltage divider was used to implement a programmable LCD reference driver
(PLRD) which takes two input arguments, a lower bound and an upper bound, as
guidance in modifying the voltage to control the transmissivity of an LCD panel.
The relation between a luminance function (bt(x)) and those two bounds is shown in
Figure 2.4, where x is pixel value, t is the transmissivity of a pixel value, and b is
a backlight factor. The luminance function consists of three regions: the undershot
region [0,gl], the linear region [gl,gu], and the overshot region [gu,1]. In other words,
the lower bound gl and upper bound gu are the darkest and the brightest pixel values
that can be displayed without contrast distortion (overshooting or undershooting)
after applying CBCS. The contrast fidelity function is defined as the derivative of
bt(x) as shown in Equation 2.2.
f
c
(x) =




















0, 0  x < gl
c, gl  x  gu, 0  c  1
1, gu < x  1
(2.2)
where, ‘c’ is limited between ‘0’ and ‘1’. If ‘c > 1’, the contrast increases and deviates
from that of the original image and the dynamic range ‘[gl,gu]’ shrinks.
The principle of CBCS is to scale the brightness and contrast simultaneously to
balance the contrast loss and the number of saturated pixels. As shown in Figure 2.5
23
Figure 2.4. Luminance as a function of Backlight and Transmissivity
simultaneous scaling both brightness and contrast presents better image fidelity than
scaling just one of these attributes. The goal is to find the optimal bounds where the
overall contrast fidelity reaches its maximum. A large overall contrast fidelity (almost
or the same as 1, which is the original image contrast) is better. The authors of CBCS
method found the optimum bounds by using this in a number of experiments. From
their experiments, they claimed that CBCS can achieve a significant power saving of
more than 50% with small contrast distortion for still images.

However, CBCS cannot maximize the potential of a dynamic backlight scaling
scheme in saving power due to its overestimation in measuring distortion [3]. CBCS
maximizes only the number of preserved pixel values or minimizes the number of
saturated pixels.
HEBS - Histogram Equalization for Backlight Scaling
Iranli et al. [3] proposed histogram equalization technique for backlight scaling
with a pre-defined distortion level. HEBS tries to find an appropriate pixel trans-
formation function for each displayed image. They argued that the image distortion
should be considered as a complex function of visual perception, and it should be mea-
24
(a) Original (b) 50% Contrast
(c) 50% Brightness (d) 50% CBCS
Figure 2.5. Visual Effects of Adjusting Brightness (b), Contrast (c), and Both (d)
when the Backlight is Dimmed to 50%
25
sured by combining the mathematical differences between pixel values (histograms)
and the characteristic of a Human Visual System (HVS). HEBS works on an image
histogram and a transformation function which transform an original histogram into
a new uniformly distributed one with a specified minimum dynamic range. Dynamic
range is a ratio or range between the brightest and the darkest available pixel values
in an image. A new histogram should be different minimally from the original one
of a backlight-scaled image. After they get the histogram transformation function,
they define a dynamic linear function to transform an original image to a desired
resultant one. Practically, since it is difficult to measure a distortion degree, this
technique needed a number of experiments to get a mapping table of dynamic ranges
to distortion ratios from some benchmark images. An example of a mapping result
is shown in Figure 2.6. The results are then used to specify the minimum dynamic
range in a new uniform distribution histogram.
HEBS method results in about 45% power saving with an effective distortion
rate of 5% and 65% power saving for a 20% distortion rate. This is significantly

higher power savings compared to previously reported ambient independent backlight
dimming approaches.
HVS Based Dynamic Tone Mapping
All of the aforementioned techniques rely on the luminance values of pixels of
the displayed image as their optimization variables. They do not consider the HVS
perceived quality. Luminance value of a light source is not the same as its perceived
brightness. Figure 2.7 shows the relation between luminance and perceived bright-
ness. The slope of each curve represents the luminance contrast sensitivity of human
26
Figure 2.6. Luminance as a function of Backlight and Transmissivity [3]
eyes, that is, sensitivity of the HVS brightness perception to the changes in the lumi-
nance. As luminance adaptation level of human eyes decreases (each curve represents
different luminance adaptation level), the luminance contrast sensitivity decreases. It
is also observable from the figure that, the HVS exhibits higher sensitivity to changes
in luminance in the darker regions of an image. Iranli et al. [36] considered the
HVS characteristics and proposed a more efficient way of backlight scaling using tone
mapping. Their method is known as Dynamic Tone Mapping (DTM) method. Tone
mapping is a classic photographic task of mapping of the potentially high dynamic
range of real world luminance values to the low dynamic range of the photographic
print. The success of photography has shown that it is possible to produce images
with limited dynamic range that convey the appearance of realistic scenes. This is
fundamentally possible because the human eye is sensitive to relative, rather than
absolute, luminance values. Many researchers have worked on automatically using
27
Figure 2.7. Luminance vs Perceived Brightness
tone mapping to compress the high dynamic range of images to a lower dynamic
range [45–48]. In addition, the benefits of different tone mapping operators have
been compared using both subjective evaluations [49, 50] and SSIM based objective
evaluations [37].
Iranli et al. [36] define luminance scaling problem as a Converse Tone Mapping

(CTM) problem and provide a possible solution to the problem. The CTM problem
is given below.
Let L
orig
max
and L
DT M
max
denote the maximum luminance of the original image and
the dynamically tone-mapped and backlight-scaled image, respectively. Moreover, let
χ
orig
and χ
DT M
denote the pixel value information of the original and backlight scaled
images. Then, the perceived image distortion between images χ
orig
and χ
DT M
can be
quantified by function D(χ
orig
, χ
DT M
).
28
Converse Tone Mapping (CTM) Problem: Given an original image χ
orig
and maxi-
mum allowable image distortion D

max
, find the tone mapping operation ψ : [0, L
orig
max
] →
[0, L
DT M
max
] such that, L
DT M
max
is minimized while,
D(χ
orig
, χ
DT M
)  D
max
, where, (2.3)
χ
DT M
≡ ψ(χ
orig
)
The proposed method (Equation 2.3) is amenable to highly efficient hardware
realisation because it does not need information about the histogram of the displayed
image. However, to realise the new tone mapping methods or operators in smart-
phones in an efficient way, significant changes are required at hardware level.
Backlight Local Dimming for LED backlit LCDs
In the previous sections, we have presented several backlight dimming techniques

which dim all the light sources together with the same dimming factor. Those ap-
proaches belong to a kind of backlight dimming scheme called a global backlight
dimming scheme. There are backlight dimming techniques which employ a different
approach, where they can dim each light source individually with different dimming
factors. These approaches belong to a local backlight dimming scheme. Shrirai et
al. [51] describe 0D, 1D and 2D adaptive dimming approach for LED backlit LCD dis-
plays. 0D dimming refers to overall dimming of the backlight (same as the dimming
technique for CCFL backlit LCD displays, described above), 1D dimming refers to
29
line dimming (linear strip of backlight) and 2D refers to individual point light source
dimming and this type of dimming is not possible in CCFL of backlight.
Dynamic Backlight Adaptation for Videos
Pasricha et al. [32] claim that aggressive compensation in luminance scaling in-
troduces noticeable artifacts in still images, but these are less discernible in video
because several frames appear on the screen every second. They propose a proxy
based architecture for dynamic backlight scaling of videos for handheld devices. All
communication between the handheld devices and the video server passes through
the proxy server, which changes the video stream in real time. The authors have
introduced the concept of a Group of Scenes (GOS) which defines the granularity
at which backlight compensation is performed. Generally, in video streams, many
frames with similar average luminosity values are clustered together, providing ample
scope to optimize for low power by uniformly compensating entire GOS entities and
reducing the handheld’s backlight level. GOS is defined as a group of contiguous
frames in a video stream such that the variance of the average luminosity values of
each frame belonging to the group is less than a threshold value.
Pasricha et al. provide three middleware adaptation policies that use the compen-
sation algorithm. The first one is Simple Backlight Compensation (SBC), in which
the backlight is reduced for GOSs which have average luminosity value greater than
the threshold (τ). The proxy sends a control signal with the GOS to the handheld
stating the amount of backlight to be reduced. The second approach is to reduce the

backlight to a prefix value and then compensate it by increasing the brightness of
the GOPs. This approach is called Constant Backlight with Video Luminosity Com-
30
pensation (CBVLC). The final approach is Dual-Compensation Approach (DCA),
in which the video stream and the backlight levels for different GOS entities are si-
multaneously compensated. The proxy dynamically compensates the GOS entities
in the video stream and begins streaming the video to the client, simultaneously
directing the client, through the control stream, to change its backlight level. The
client middleware sets the appropriate backlight intensity levels for the video play-
back. This approach provides more flexibility for aggressive optimizations and results
in far greater power savings. In addition, the video frames are convolved with a
high pass filter to minimise impact on picture detail (loss of contrast) after aggressive
luminosity compensation.
However, Pasricha et al. focus only on luminosity compensation, ignoring the
characteristics of the HVS. Moreover, applying linear changes to the pixels of each
frame is computationally intensive for real-time video.
Quality Based Backlight Adaptation for Videos
Cheng et al. [35], employ a technique to incorporate video quality into the back-
light switching strategy and proposes a Quality Adaptive Backlight Scaling (QABS)
scheme. The backlight dimming affects the brightness of the video. Therefore, QABS
only consider the luminance compensation such that the lost brightness can be re-
stored.
Reducing backlight and enhancing image to compensate the loss in brightness,
results in quality distortion due to clipped pixels. The authors use Mean Square Error
(MSE) (Equation 2.4) to measure the quality distortion and control the backlight
dimming based on MSE threshold (Q
th
).
31
Figure 2.8. Relation Between MSE and Backlight Level

MSE =
1
M
×
M

i=1
(x
i
− y
i
)
2
(2.4)
where, ‘x
i
’ and ‘y
i
’ are the original pixel value and the reconstructed pixel value,
respectively. ‘M’ is the number of pixels per frame.
The dynamic adaptation algorithm takes a threshold value Q
th
and performs an
exhaustive search on the model shown in Figure 2.8 to find the optimal backlight level,
Alfa. Then, the backlight is set to Alfa and the loss of brightness is compensated by
increasing image illuminance. To reduce the frequency of backlight switching, they
propose two supplementary schemes to smooth the backlight switch process such
that the user perception of the video stream can be substantially improved. First, a
low-pass digital filter is proposed to eliminate any abrupt backlight switching that is
caused by the unexpected sharp luminance change. Second, they propose to quantize

32
the number of backlight levels, that is, any backlight level between two quantization
values can be quantized to the closest level, by which we prevent the needless backlight
switching for small luminance fluctuations during one scene. The evaluations show
that by applying the scheme, up to 40% power can be saved with negligible loss to
video quality. However, the metrics used for distortion and quality do not consider
the characteristics of HVS. It is well known that MSE and PSNR (Peak Signal to
Noise Ratio) are not the best measures to assess perceptual quality for most video
sequences [52] [53]. Widely adopted metrics such as, SSIM (Structural Similarity
Index) [54] can be used provide better estimation and quality.
2.2 OLED Display Power Conservation
The power consumption of an OLED displays depends on the Lightness and Colour
(Hue) of the pixels displayed. Energy consumption of an OLED display can be com-
puted as sum of energy consumption of all pixels. Energy consumption of a pixel is
simply sum of energy used to illuminate R (Red), G (Green) and B (blue) organic
materials which forms the pixel [55, 56]. We confirmed this with our own mesure-
ments and presented the results in Section 4.1 Chapter 4. Hence transforming the
content colours to their power efficient versions has become the primary mechanism
for conserving power in OLED devices. There are several works on colour transfor-
mation.
Energy Aware Colour Set
Chuang et al. [4] propose two colour mapping approaches that swaps current
colours with an iso-lightness colour set within a certain restrictions, so that the overall
33
lightness of the final display is still the same with the original display. The first
one is discrete optimisation approach, in which, from a set of M iso-lightness and
distinguishable colours, a subset of N colours (N ≤ M) is chosen so that the sum of
the energies associated with the chosen colours is the minimum of any subset of N
colours. To select the set of M adequate input colours the authors propose to adopt
named (categorical) colours because they are sufficiently distinct and they exhibit

proven perceptual benefits. For example, Kawai et al. [57] suggest that the time
it takes to distinguish multiple colours depends partly on their named colour region.
The number of easily distinguishable colours is small. For example, in his study
of categorical colours, Healey [58] finds that 7 distinct iso-lightness colours is the
maximum number of colours that can be displayed at one time without lowering the
response time and accuracy of target colour identification. Based on Healey’s [58]
study, Chuang et al. have selected 6 quickly identifiable colours (green, blue, orange,
purple, red, yellow) measured their power consumption (Figure 2.9 and Figure 2.10).
L
∗
in these figures indicates lightness parameter as defined in CIELAB color space
and E denotes normalised energy consumption. Users can pick distinguishable iso-
lightness colours with increasing energy cost by choosing colours from bottom to top
from the figures.
The second one is continuous optimisation approach, in which, the goal is to
find energy aware, distinguishable, iso-lightness colours with 3 input parameters: the
number of colours N, the level of lightness L
∗
, and a minimum perceptual colour
distance d that must be enforced between every pair of colours.The rationale is that
this minimum distance is a means of ensuring distinguishability of the colours. L
∗
is
34
Figure 2.9. Energy plot of the hues of 6 categorical colours (from left to right: blue,
red, purple, orange, green, yellow). Evergy (E) vs Lightness (L
∗
) [4] .
Figure 2.10. Categorical colours of varying lightness sorted by increasing energy
cost [4] .

35
(a) Original Colourset (b)Discrete Optimization (c) Continuous Optimization
Figure 2.11. Tooth Dataset Coloured with Traditional Colours (Original) and En-
ergy Efficient Colours [4]
the lightness parameter as defined in CIELAB ( International Commission on Illumi-
nation recommended L
∗
, a
∗
, b
∗
) colour space. A cost function is defined that takes
into account the energy of the colours and the distances between them in selecting
the colours.
In both approaches, colour distinguishability and iso-lightness are ensured. The
examples for 2D colour mapping and volume rendering achieves around 40% energy
saving. However in Chuang et al.’s approaches, the original colours are significantly
changed (Figure 2.11) in the process and not suitable to be applied to photos and
videos, in which colour accuracy is important.
Colour Transformation for GUIs
Dong et al. [28] describe two approaches of reducing power consumption of OLED-
based displays on mobile phones through structured colour transformation of standard
background and foreground Graphical User Interface (GUI) objects, or unstructured
transformation of individual pixel colours of arbitrary GUI elements. In unstructured
36
transformation, a power efficient colour for background is identified first, then colours
which have high contrast with the selected background colour are be used as window
borders and foreground colours. In unstructured transformation, Dong et al. do a
exhaustive search in the colour space to replace each colour in the GUI with a power
efficient colour. To ensure readability, the approach ensures the colour difference

between every two colours in the transformed GUI is no less than their counterparts
in the original.
Dong et al. propose two options for unstructured transformation. First one is
mono-colour mapping where the colours of the GUI are mapped to a single colour
with varying lightness. The second one is rank-based colour mapping where colours of
the GUI are mapped to different set of colours which are power efficient. In their study,
they observed that users preferred mono-colour than multi-colour GUI. Their imple-
mentation takes power efficient requirement and options (structured/unstructured;
mono-colour/multi-colour) as input and maps to energy efficient colours. However,
the authors did not consider one of the key property (Colour Harmonicity) that deter-
mines the visual quality of the GUI. This may make the GUI visually unpleasant for
the human eyes. In addition, the approach dramatically alters the original colour of
the displayed objects (shown in Figure 2.12), again making it unsuitable for contents
which are sensitive to colour fidelity changes, such as images and videos.
Colour Transformation for Webpages
In their next paper, Dong et al. [6] presented a mobile web browser which trans-
forms the colours of the webpages to reduce power consumption. Their transforma-
tion logic, has four options, namely Dark, Green, Arbitrary and Inversion. Dark and
37
Original Green MultiColour
Figure 2.12. Unstructured Transformed GUIs with Different Settings
Green options map the colours of the web page to darker version or shades of green
colour as green being the most energy efficient colour. However, this makes the entire
web pages look dark and greenish. Moreover, it won’t be interesting to view all web
pages in same colour for the users. This may bring dissatisfaction to the owners or
designers of the web sites. Arbitrary approach maps the colours to arbitrary colours
that are energy efficient and Inversion approach simply inverts the colours.
In all the four options, the brand identity (brand colour) of the page is lost making
the page not recognisable by its colour scheme. For example, their results for Enter-
tainment and Sports Programming Network (ESPN) webpage (excluding Darkening

option) shown in Figure 2.13 has totally different colours than the brand colours of
ESPN, which are red, white and black. The foreground images are simply darkened
without considering HVS characteristics. Applying pixel-by-pixel colour transforma-
tion during the rendering phase of the browser is also compute-intensive task. An
alternative processing, such as, processing the source code of the web page for colour
transformation will be computationally efficient.
38
Original Green
Arbitrary Inversion
Figure 2.13. Colour Transformed ESPN Webpage
39
Original Colour Quantized
Figure 2.14. Colour Quantizing an Image to N = 32 Colours
Just Noticeable Deference Constrained Transformation Scheme
Hadizadeh et al. [59] describe a colour mapping algorithm which maps the image
colours to power efficient versions constrained by Just Noticeable Deference (JND)
threshold. The authors first colour quantize the entire image in to set of N distinct
colours (C
1
, C
2
C
N
) as shown in Figure 2.14. The colour of all pixels in a quantum
i is same (C
i
). Then, each colour C
i
is replaced with another colour which is energy
efficient and lies within the JND threshold.

To obtain energy efficient colours, the authors propose the following algorithm.
Let the colour of the given pixel C be (Y, Cb, Cr) in Y CbCr colour space. They
first compute JND
Y
which is the spatial luminance JND computed from the Y
component of the colour C. Adding or subtracting upto JND
Y
amount of luminance
to the Colour C

s luminance component Y will have unnoticeable effect for human
eyes. Hence, two new colours, C
+
and C
−
are created, by adding and subtracting
JNDy. These two new colours can be considered visually indistinguishable from C
for human eyes, since their chroma components are the same as those of C, and
the difference between their luminance components and the luminance component
of C does not exceed the JND threshold. The three colours (C, C
+
, C ) are then
40
transformed to CIELAB colour space, and the CIEDE2000 distances [60] between
them are calculated (Equation 2.5).
R
+
= D
00
(C, C

+
)R = D
00
(C, C) (2.5)
Due to the nonlinear transformation from YCbCr to CIELAB, R
+
may be different
from R. Hence, minR
+
, R is set to R. Now, all colours in CIELAB whose distance
D
0
0 from C does not exceed R should be perceptually indistinguishable from C.
Among all these colours (that are within distance R to the colour C), the one with
minimum energy consumption is selected for replacing the colour C.
In reality spatial luminance JND of most of the colours are very low. Hence, it
results in very small value of R. This results in image quality which is close to original
(Figure 2.15). However, due to the strict JND constraint, this scheme saves lessthan
5% of display energy which translates to 2% of system energy. If we consider the
computational costs, this scheme becomes inefficient and may even result in overall
negative energy saving.
Log-modified Histogram Equalisation Scheme
In an image intensity histogram with a column vector h, the k-th element h
k
denotes the number of pixels with intensity k. Histogram equalisation process or
function maps input pixel intensities to output pixel intensities, to make the his-
togram of the output image as uniform as possible. Large values of histogram bins
affects the HE function. Hence, pre-modification is applied before HE. Chulwoo et
al. [38] propose an alternative histogram pre-modification approach, which is log-
based and reduces large values of histogram bins before the HE procedure to avoid

41
Original Colour Quantized (N=512) Power Optimised
Figure 2.15. Colour Quantizing an Image to N = 512 Colours
extreme slopes in the transformation function. The authors describe a Log-Modified
Histogram Equalisation (LMHE) scheme, which reduces over stretching artifacts of
the conventional histogram equalisation technique. In addition, the algorithm is con-
strained by OLED power model. Effectively their scheme, enhances the contrast by
equalizing the histogram, and saves the power consumption by reducing the histogram
values for large intensities.
However, their approach is more focused on contrast enhancement than power
saving. They compare the results of their HE algorithm with other HE algorithms.
However, they do not provide sufficient evaluation on the quality of their power
optimised images. The power saving measurements shown in Figure 2.16, depicts
that their LMHE approach (without power constraint α = 0) does not save significant
amount of energy when compared to conventional HE algorithm. α is a parameter
to their transformation function which determines the weightage for power efficiency.
Though the approach saves power when α > 0, the image gets darker (Figure 2.17)
42