D046022629

Paramjit Singh Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 6( Version 2), June 2014, pp.26-29
www.ijera.com 26|P a g e
Contrast Enhancement of Color Images with Bi-Histogram
Paramjit Singh, Dr. Amandeepkaur
(Department of Computer Science, Punjabi University Patiala)
ABSTRACT
Histogram equalization is a widely used scheme for contrast enhancement in a variety of applications due to its
simple function and effectiveness. One possible drawback of the histogram equalization is that it can change the
mean brightness of an image significantly as a consequence of histogram flattening. Clearly, this is not a
desirable property when preserving the original mean brightness of a given image is necessary. Bi-histogram
equalization is able to overcome this drawback for gray scale images. In this paper, we explore the use of bi-
histogram equalization based technique for enhancing RGB color images. The technique is based on cumulative
density function of a quantized image. From the results it is concluded that bi-histogram equalization is able to
improve the contrast of colored images significantly.
Keywords— histogram equalization, histogram specification, Image enhancement, maximum entropy, mean
brightness preserving.
I. INTRODUCTION
Histogram equalization is a well-known method
for enhancing the contrast of a given image in
accordance with the sample distribution. In general,
histogram equalization flattens the density
distribution of the resultant image and enhances the
contrast of the image as a consequence [4]. In spite of
its high performance in enhancing contrasts of a
given image, however, Global histogram equalization
may change the original brightness of an input image,
deteriorate visual quality, or, introduce some
annoying artifacts [ woods].
Brightness preserving Bi-Histogram
Equalization (BBHE) has been proposed to overcome
these problems [2]. BBHE first separates the input
image’s histogram into two by using the image mean.
Next, it equalizes the two sub-histograms
independently. It has been analyzed that BBHE can
preserve the original brightness to a certain extent
when the input histogram has a quasi-symmetrical
distribution around its mean. Later, equal area
Dualistic Sub-Image Histogram Equalization
(DSIHE) has been proposed [6], it claims that if the
separating level of histogram is the median of the
input image’s brightness, it will yield the maximum
entropy after two independent sub-equalizations [1].
Another scheme, named Recursive Mean-Separate
Histogram equalization (RMSHE), has been
proposed to preserve the brightness. RMSHE uses the
BBHE iteratively. It is claimed theoretically that
when the iteration level n grows larger, the output
mean converges to the input mean, and thus yields
good brightness preservation [7].
II. HISTOGRAM EQUALIZATION
Histogram equalization can be categorized into
two methods: global and local histogram
equalization. Global histogram equalization uses the
histogram information of the whole input image as its
transformation function. This transformation function
stretches the contrast of the high histogram region
and compresses the contrast of the low histogram
region. In general, important objects have a higher
and wider histogram region, so the contrast of these
objects is stretched. On the other hand, the contrast of
lower and narrower histogram regions, such as the
background, is lost.
This global histogram equalization method is
simple and powerful, but it cannot adapt to local
brightness features of the input image because it uses
only global histogram information over the whole
image. This fact limits the contrast-stretching ratio in
some parts of the image, and causes significant
contrast losses in the background and other small
regions. To overcome this limitation, a local
histogram-equalization method has been developed,
which can also be termed block-overlapped
histogram equalization. [1]
Digital image with gray levels in the range [0, L
−1], Probability Distribution Function of the image
can be computed as equation,
P(rk)=
𝑛 𝑘
𝑁
k=0,…….L-1 (1)
Where rk is the kth
gray level and nk is the
number of pixels in the image having gray level rk.
Cumulative Distribution Function (CDF) can also be
computed followed,
C(𝑟𝑘 ) = 𝑃(𝑟𝑖)𝑖=𝑘
𝑖=0 (2)
K=0,…….,L-1, 0 ≤ C(rk) ≤ 1 (3)
Let X={X(i,j)} denotes a given image composed of L
discrete gray levels denotes as {X0,X1 ,……XL-
RESEARCH ARTICLE OPEN ACCESS

1},where X(i,j) intensity of the image at the spatial
location (i,j) and X(i,j)Є{X0,X1, XL-1}. In image X,
the probability density function p(Xk ) is defined as,
p(Xk)=
nk
n
(4)
In which k=0,1…….L-1, where nk
represents the
number of times that the levels Xk appears in the
input image X and n is the total number of samples in
the input image. A plot of nk
and XK is known as the
histogram of X, based on the probability density
function, we define the cumulative density function
as,
C(x) = pk
j=0 (Xj) (5)
In this equation Xk =x for k=0,1…..L-1.
Histogram equalization is a scheme that maps the
input image into the entire dynamic range, (X0, XL-1)
by using the cumulative density function as a
transform function. That is, let ns define a transform
function f(x) based on the
Cumulative density function as,
F(x)=X0 + (XL-1-X0) c(x) (6)
then the output image of the histogram equalization,
Y = {Y (i,j ) } , can be expressed as,
Y=f(X)
={f(X(I,j))∀X (i, j)X} (7)
III. BI HISTOGRAM
EQUALIZATION(BHE)
Bi Histogram Equalization is used for improved
histogram equalization (HE) method for contrast
enhancement.[5] BHE first finds average point in
histogram of the image and then divides histogram to
two segments based on this point. After that
histogram equalization operation is applied on each
segment. There are two cumulative distribution
functions for two segments. Gray level ( rk ) under
the average point are pointed to the new gray level (
Sk ) as it can be seen in equation.[2]
Sk = (L1) × C1(rk)(8)
C1(rk) = ( 𝑛𝑖
𝑘
𝑖=0 ) / ( 𝑛𝑗
𝑘
𝑗=0 ); (9)
K=0,1,……,L-1
Where L1 is the average of gray levels of the
histogram and it can be computed as in equation. [5]
L1 = 𝑃 𝑟 𝑘
𝐿−1
𝑘=0 ∗ 𝑟𝑘 (10)
IV. DUALISTIC SUB-IMAGE
HISTOGRAM EQUALIZATION
DSIHE [5] is very similar to BBHE, except that
the separating point XDis selected as the median gray
level of the input image, i.e., XDsatisfies,
pr r dr = 0.5
D
0
(11)
For the applicable case, it may be modified as,
XD=argminx | pr r dr − 0.5
x
0
(12)
The purpose of DSIHE is to find a separating
point, based on which the desired histogram (PDF)
can obtain a maximum entropy. And it has been
proved that the brightness of the output image is,
µz= 0.5 (XD+ 0.5) (13)
It is clear that DSIHE always pulls the output
brightness toward the input middle level from the
input median level.
V. RMSHE and MMBEBHE
RMSHE [6] is to recursively implement BBHE to the
histogram. Obviously, recursive separation on the
histogram will divide the histogram into very small
pieces. If the recursive level tends to infinite,
equalization to each small piece will not change the
whole histogram, and thus can preserve the mean
brightness. So an infinite recursive level results in the
same output image as the input one. BBHE and
DSIHE belong to “two-piece-separating” histogram
equalization algorithm, and so does HE in a broad
sense.
ProposedScheme forContrast Enhancement of
Color Images
I. Read color image and find mean of image
II. Find two histogram of input image one is
xlpdf=hist(xl,[0:xm]); (14)
Second is xupdf=hist(xu,[xm+1:255]); (15)
III. Apply the code for enhance the image
skl=xm*xlnpdf*triu(ones(xm1));. (16)
sku=(xm+1)+((L-1
(xm+1))*xunpdf*triu(ones(xm2))); (17)
IV. Find the non-zero values
list=find(I==k);
V. Increase the list of value
y0(list)=sk(k+1);
VI. Results and Discussion
The proposed method was tested using different
colored images. For testing a dataset is generated by
taking good contrast images and saving them as
benchmarks, after that their contrast is reduced and
these low contrast images are enhanced using the
proposed algorithm. The result was evaluated both
qualitatively and quantitatively. The qualitative
results compared with the benchmark images are
shown in the Fig. 1 for two different images.
The results of the proposed algorithm for two test
images are shown in Fig 1. As is evident from the

output images (E, F) their contrast is significantly
improved from low contrast images (C, D) and even
from the benchmark image (A, B).
Figure 1: original images shown as A, B. Images
after decreasing their contrast C, D. Results of the
proposed algorithm for the three images E, F.
For quantitative measures peak signal-to-noise
ratio (PSNR) is used to evaluate the results of the
proposed algorithm. The PSNR, is an engineering
term for the ratio between the maximum possible
power of a signal and the power of corrupting noise
that affects the fidelity of its representation. Because
many signals have a very wide dynamic range, PSNR
is usually expressed in terms of the logarithmic
decibel scale.
The PSNR is defined as:
PSNR = 10. log10
MAX 2
MSE
(18)
Table 2 gives PSNR values. PSNR values of the
enhanced images are significantly imposed by
applying Bi-histogram recursively to low contrast
images. On average is 25% improvement from
original low contrast images
Table 2: PSNR values of low contrast and enhanced
images
No. Origin
al
Image
Low
Conrta
st
Bi-
Histogra
m
Recursi
ve Bi-
Hist
1 A 14.5287 20.1598 42.1489
2 B 17.6055 19.2628 36.8856
3 C 13.6470 20.6009 40.9992
4 D 12.2553 22.0261 32.2151
5 E 15.7666 23.4272 41.2501
The role of different color channels in RGB
images is also analyzed. In experiment two channels
are kept the same and only one channel is enhanced
before combining the three to get the output RGB
image. The results are shown in Figure 2 and Table 3.
Figure 2 shows the visual results for one test image.
The results show that modification in the Red
channel gives the highest PSNR values for test
images. Hence we may conclude that enhancement is
more sensitive to Red channel.
Low contrast Image
Enhanced using Red channel
Enhanced using Green channel
Enhanced using Blue channel
Fig 2: Test image A and its enhancement using
individual color channels (RGB)
Table 5.3: PSNR values after enhancing single color
channel
Test
Images
Red Blue Green
1 26.9683 26.1012 24.7322
2 27.2416 26.5309 29.1843

VII. CONCLUSION
In this paper, a bi-histogram algorithm used in
gray scale images is modified for colored image
contrast enhancement. The proposed algorithm gives
promising results. The other advantage of using this
method is its simple implementation and low time
complexity.
REFERENCES
[1] R.C. Gonzalez and R.E. Woods “Digital
image processing”, third edition, 2008
[2] M. Yeong-Taekgi, “Contrast Enhancement
Using Brightness Preserving Bi-Histogram
Equalization”, IEEE Transactions on
Consumer Electronics, vol. 43, no. 1, February
1997
[3] R. Anoop and M. Wilscy, “Enhancement
contrast in color images using bilateral filter
and Histogram Ezualization using wavelet
coefficient”, 978-0-7695-3546-3/08, 2008
[4] H. Yeganehl, A. Ziaei, A. Rezaie, “A Novel
Approach for Contrast Enhancement Based on
Histogram Equalization”, Proceedings of the
International Conference on Computer and
Communication Engineering 2008 May 13-15,
2008
[5] A. Phillip, Z. Qiang, and J. JeflreyRodrz, “3-D
Histogram Modification of Color Images”,
Proc. IEEE, pp. 1015-1018, 1996, pp. 1015-
1018
[6] S. P. K.Nicholas, I. Haidi, “Color Image
Enhancement Using Brightness Preserving
Dynamic Histogram Equalization”, IEEE
Transactions on Consumer Electronics, vol.
54, no. 4, November 2008
[7] C. Soong-Der., R. Rahman, “Minimum Mean
Brightness Error Bi-Histogram Equalization
in Contrast Enhancement”, IEEE Transactions
on Consumer Electronics, vol. 49, no. 4,
November 2003

D046022629

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to D046022629 (20)

Recently uploaded (20)

D046022629