Sie sind auf Seite 1von 6

A New Approach To Image Contrast Enhancement

using Weighted Threshold Histogram Equalization


with Improved Switching Median Filter

R.Sharmila
Rajiv Gandhi College of Engineering and Technology
Electronics and Communication Engineering
Puducherry, India
Sharmila1ece@gmail.com
R. Uma
Rajiv Gandhi College of Engineering and Technology
Electronics and Communication Engineering
Puducherry, India
Uma.ramadass1@gmail.com


Abstract Image contrast enhancement plays a vital role in
digital image processing especially in biomedical
applications and secures digital image transmission. The
objective of image enhancement technique is to improve
the characteristics or quality of an image, such that the
resulting image is better than the original image. To
improve the image contrast, numerous enhancement
techniques have been proposed. One of the conventional
methods adopted is the Histogram Equalization (HE)
technique. The main limitation of this technique causes
unpleasant visual artifacts, such as over enhancement,
level saturation and raised noise level. To overcome these
limitations technique like Dualistic Sub-Image Histogram
Equalization (DSIHE), Brightness Preserving Bi-
Histogram Equalization (BBHE) and Weighted Threshold
Histogram Equalization (WTHE) has been proposed
which failed to eliminate the impulse noise. In this paper
we proposed a novel method called Weighted Threshold
Histogram Equalization with Improved Switching Median
Filtering technique (WTHEISMF) which eliminates the
impulse noise and preserves the edges of an image. This
paper presents the parameters comparison and simulation
results of HE, WTHE and WTHEISMF with Absolute
Mean Brightness Error (AMBE), Measure of
Enhancement (EME), Peak Signal to Noise Ratio (PSNR),
Mean Square Error (MSE).

Keywords- Histogram Equalization (HE), contrast
enhancement, Impulse noise, Improved Switching Median
Filter (ISMF)
I. INTRODUCTION
The image enhancement is to improve the interpretability of
the information present in images for human viewers. An
enhancement algorithm is one that yields a better quality
image for the purpose of some particular application which
can be done by either suppressing the noise or increasing the
image contrast. Image enhancement algorithms are employed
to emphasize, sharpen or smoothen image features for display
and analysis. Image enhancement techniques can be classified
into two broad categories they are Spatial domain method and
Transform domain method. The spatial domain method
operates directly on pixels, whereas the transform domain
method operates on the Fourier transform of an image and it is
transformed back to the spatial domain. There are two
techniques in contrast enhancement: Direct methods and
Indirect method [1][2]. There are several subgroups in Indirect
methods[3] i) techniques that decay an image into high and
low frequency signals for manipulation ii)histogram
modification techniques and iii)Transform based techniques.

Elementary enhancement techniques are histogram based
because they are simple, fast and produces acceptable results
for some applications can be achieved. Histogram
modification basically modifies the histogram of an input
image so as to improve the visual quality of the image.
Histogram equalization is a process that attempts to spread out
the gray levels in an image so that they are evenly distributed
across their range. In the histogram equalization without any
modification, it over enhanced the output image and raised
noise level. A number of methods have been proposed [3] for
limiting the level of enhancement by modifying the histogram
equalization.

The recent method proposed to enhance an image using HE is
Adaptive HE methods and Global methods based on
Histogram equalization [4]. In the above two proposed
method, AHE methods can provide high level of enhancement
compare to the Global methods. Bi-Histogram Equalization
was proposed to reduce the mean brightness change in an
image. BHE is a modified version of Histogram Equalization
[3, 4]. In this method initially an image is partitioned into two
sections based on average of the histogram and then applies
histogram equalization on each section individually. A similar
improved Histogram based method is Dualistic Sub-Image
Histogram Equalization (DSIHE)[3,4]. In these methods,
initially histogram is partitioned into two fragments based on
entropy and then applies histogram equalization on both the
fragment. In the above method the enhanced image is more
satisfying than Histogram equalization but not possible to
I
J
A
E
S
T
R.Sharmila* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 2, 206 - 211
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 38
change the level of enhancement [3]. The aforementioned
problems can be overcome by Weighted Threshold Histogram
Equalization.

The general idea adopted by the WTHE [4] method is to
modify the histogram before equalization is conducted. Such
modifications reduce visual artifacts. During the image
enhancement the impulse noise in the image also enhanced
[5]. To avoid this effect, the enhanced image is passed through
a filter. In digital signal and image processing; an image is
often corrupted by noise in its acquisition or transmission.
Noise is undesirable information that contaminates an image.
So the conventional filtering techniques are used to eliminate
the noise and to preserve the edges. Several filtering
techniques have been reported are linear filtering; averaging
filtering and nonlinear filtering [6]. The linear filtering tends
to blur the edges, and do not remove the noise effectively.
Nonlinear filtering is generally used for the removal of
impulse noise due to some advantages like preservation of
edges. The nonlinear filtering is used in various fields like
image processing, Telecommunication, Geo physical signal
processing.

The median filter is an effective method for the removal of
impulse based noise on images [5]. It is also called as sliding
window filter, which replace the center value in the window
with the median of all pixel values in the window. This is due
to the partial averaging effect of the median filter and its
biasing of the input stream, rather than straight mathematical
averaging. The drawback of using this median filter is that
they cannot detect whether the pixel is noisy or not, so they
replace both noisy as well as noise free pixel which makes the
image to become visible blur. The aforesaid problems can be
overcome by modified forms of the median filter have been
proposed [5].The weighted median filtering method assign
weight to each pixel values; it duplicates the pixel values for
several times. The existing median filtering technique
decreases the quality of the enhanced images rather than
improving the quality of an image. The Improved Switching
Median Filtering technique [7] is used to reduce the noise
level in enhanced image and preserving the edges of an image.
So this paper presents a new novel approach for image
contrast enhancement using Weighted Threshold Histogram
Equalization with Improved Switching Median Filter.

This paper is organized as follows. Section II gives the
Histogram Equalization technique procedure. Following
section III explains the details about Weighted Threshold
Histogram Equalization. Section IV presents the Improved
Switching Median Filtering technique and its advantage. In
section V presents the detail about proposed method,
Simulation results and discussions are presented in Section VI.
Finally Conclusion is provided in section VII. The simulation
result is tested using MAT Lab tool.

II. HISTOGRAM EQUALIZATION TECHNIQUE
The main objective of histogram equalization is to obtain a
uniform histogram for the output image [3]. The procedure to
perform histogram equalization techniques is as follows:

1. Find the running sum of the histogram values.
2. Normalize the running sum values by dividing by the total
number of pixels.
3. Multiply the normalized values by maximum gray level and
round off to the nearest integer value.
4. Map the gray level values using a one to one mapping.

For a given digital image G (i,j) with M pixels and a gray level
of [0,L-1].The probability density function(PDF) is given by
( )
n
n
u
P k
M
= for 0,1,....., 1
n
k L = (1)
Where is
n
u the number of times level k appears in an image.
The running sum or cumulative density function (CDF) of
histogram values is obtained by
0
( ) ( )
k
n
m
C k P m
=
=

for 0,1,....., 1
n
k L = (2)
Using the running sum, histogram equalization plots an input
level
n
k into an output level
n
k
( 1) ( )
n n
k L C k = (3)
The output level can be increment by
( 1) ( )
n n
k L P k A = (4)
The above equation indicates that distance between
n
k and
n
k +1 has direct relation with PDF of the input image at the
gray level
n
k . In the Histogram Equalization, the pixel values
are uniformly distributed with gray level values across their
range. HE over enhances the background of the image and
often produces the improbable effects in the image.

III WEIGHTED THRESHOLD HISTOGRAM
EQUALIZATION
One of the recent methods proposed to overcome the visual
artifacts of the HE method and add more flexibility to it is
Weighted Threshold Histogram Equalization [4]. The general
idea adopted by WTHE is to modify the histogram before
equalization. Weighted Threshold Histogram Equalization is
performed by modifying the histogram of an image by
assigning the weight and threshold to each pixel before
equalization.

The weighted and threshold value of probability density
function ( )
n
P k is obtained by the following equation given
by,


I
J
A
E
S
T
R.Sharmila* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 2, 206 - 211
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 39




(5)





Where ( )
wt n
P k is the weighted and threshold parameter value
of ( )
n
P k .
u
P is upper threshold value, that is the PDF is
clamped to maximum gray level value of ( )
n
P k .In the
WTHE method the value of
u
P is given by

max
.
u
P v P = 0<v s1 (6)
Where
max
P is the peak value of the ( )
n
P k
The original PDF of ( )
n
P k is clamped to an upper and lower
thresholds value by using a normalized power law function
with index 0 r > .The value of r is an important parameter
which control the degree of enhancement and depends upon
the value of r ,it gives higher and lower weights to the
probability density function. After the weighting and threshold
operation the remaining procedure is similar to Histogram
Equalization.

The cumulative distribution function (CDF) of original image
G (i,j) is given by the equation
0
( ) ( )
k
wt n wt
m
C k P m
=
=

for 0,1,....., 1
n
k L = (7)
The histogram Equalization maps an input level
n
k into an
output level
n
k is given by
( 1) ( )
n wt n
k L C k = (8)

The output level can be increment by
( 1) ( )
n wt n
k L P k A = (9)
The resultant image G (i,j) is given by

( , ) ( ( , ))
out wt adj
G i j W C G i j M = +
(10)
The parameter
out
W and
adj
M are constant and it gives
dynamic range of the output image and reimburse the mean
change after enhancement. In the WTHE, it reduces the visual
artifacts resulting from the histogram equalization and it is
capable to control the level of the enhancement.

IV IMPROVED SWITCHING MEDIAN FILTERING
TECHNIQUE
Filtering plays an important role in any signal processing
application. The main goal of filtering is to enhance a fine
detail in an image. In the improved switching median filtering
technique, first find the difference between the median value
of pixels and center pixel value in the filtering window and
then compared with a two threshold values to obtain a
presence of noise in an image. The decision based filtering
procedure is used to modify the corrupted pixels; it does not
affect uncorrupted pixels. The median filtering have some
undesirable features like 1) thin lines and other fine details are
distorted 2) it replace both the noisy as well as noise less pixel
which makes the image to appear blur. The improved
switching median filter has proved to be good because it
effectively modifies only those pixels that are considered to be
corrupted pixel. The window size may be either 3 x 3, 5x5 or
7x7. In this method, the impulse noise is obtained by using the
following algorithm,

Algorithm 1: Improved Switching Median Filter
___________________________________
1. Median value n (i, j), for every window
2. ( ( , ) ( , ) ) 1 if n i j c i j threshold <
3. min{ ( , ) ( , )};
n
a c i j c i j =
4. ( 2) if a threshol <
5. ( , ) ( , ) d i j c i j =
6. else
7. ( , ) ( , ) d i j n i j =
8. ; end
9. ; end
_______________________________________


Consider an image of dimension 255 255 on which ISMF
the window size may be either 33 or 55. First step is to
find the median value. The conventional median operation is
performed by arranging the pixel values in ascending order
and the middle value is taken as the median value n (i, j) .In
step 2, find the difference between the median value n (i, j)
and the current pixel value ( , ) c i j . In step3 set the first
threshold value and the difference value is being compared
with the threshold value. If the difference is less than the
threshold value difference between the each pixel value in the
window and center pixel value is calculated. The minimum
value a is compared with the second threshold value .If the
minimum value is less than the threshold value then the
current pixel value is considered to be noise free and retained
as such. Otherwise the current pixel is considered to be noisy
and that value is replaced by the median value. By using this
modified algorithm, improved switching median filter has
proved to be good because it effectively modifies only those
pixels that are considered to be corrupted pixel.
I
J
A
E
S
T
R.Sharmila* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 2, 206 - 211
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 40
Input Histogram
Modify the PDF OF
histogram using Weighted
and Threshold Method
Histogram equalization
Enhanced image of WTHE
Improved Switching
Median filter
Enhanced Image

V PROPOSED METHOD

In this section, a Weighted Threshold Histogram Equalization
with Improved Switching Median Filtering technique
(WTHEISMF) is presented. As mentioned in section I ,there
are some problems in histogram modification techniques .We
overcome these problems in our proposed method by the
Weighted Threshold Histogram Equalization with Improved
Switching Median Filtering technique (WTHEISMF).

The proposed method consist of two steps
1. Enhance the image using Weighted Threshold
Histogram Equalization
2. Enhanced image is passed through an ISMF























Fig. 1 Flowchart of Proposed method

The flow chart of proposed method is represented in fig.1.In
this method using the weighted threshold histogram
equalization techniques the image is enhanced, preserves the
over enhancement and level clipping. While the enhanced
image is passed through an improved switching median filter,
it reduces the impulse noise created during the enhancement.


VI SIMULATION RESULTS AND DISCUSSION

In order to test the proposed method, experiments are
performed on CT scan Brain and Face of size 256 256.To
evaluate the image enhancement and restoration performance,
Absolute Mean Brightness Error(AMBE), Measure Of
Enhancement(EME), Peak Signal To Noise Ratio(PSNR),
Mean Squared Error(MSE) used as the criterion.

The AMBE is defined as total difference between input and
output mean values [15].

( ) ( )
n n
AMBE E X E Y = (11)
The measure of enhancement (EME), it gives an average
contrast of an image by partition an image into non
overlapping segments. The measurement is based on
maximum and minimum intensity of each segment and
averaging them. The MSE value is given by
1 1
2
0 0
1
( ( , ) ( , ))
j i
m n
MSE x m n y i j
ij

= =
=

(12)
Where the size of an image is i j .The original and restored
images are ( , ) x m n and ( , ) y i j .The PSNR is the ratio
between the maximum possible power of a signal and the
power of corrupted noise.

PSNR =20log
10
(255/sqrt (MSE)).

The performance of the proposed method is compared with the
conventional histogram equalization technique followed by
median and improved switching median filter and the result of
weighted threshold histogram equalization followed ISMF is
compared with the median filter. Only a few of the results are
shown in this paper.

Fig. (2,3) shows the original images and their corresponding
contrast enhanced image using HE and WTHE followed by
median and ISMF techniques.



(a) (b)
I
J
A
E
S
T
R.Sharmila* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 2, 206 - 211
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 41

(c) (d)












(e) (f)
Fig. 2 Results for CT Scan Brain image (a) Original Image (b)
HE image (c) HE followed by Median filter (d)HE followed by
ISMF (e) WTHE Image (f) WTHE followed by ISMF (proposed
method)




(a ) (b)








(c) (d)
(e) (f)
Fig.3 Results for Face image (a) Original Image (b) HE image (c)
HE followed by Median filter (d)HE followed by ISMF (e)
WTHE Image (f) WTHE followed by ISMF

TABLE 1 QUANTITATIVE MEASUREMENT RESULTS

Image AMBE

EME
HE WTHE HE WTHE

Brain

559.2421

28.0946

9.3772

6.1649


Face

441.2384

69.9803

10.9344

5.8623


TABLE 2 .MSE and PSNR for proposed and existing method


Image MSE PSNR
WTHE
followed
by
Median
filter
WTHE
followed
by
ISMF
WTHE
followed
by
Median
filter
WTHE
followed
by
ISMF

Brain

5.7227

3.4125

40.5548

42.8000


Face

3.8017

2.9057

42.3310

43.4983


The proposed algorithm is compared with histogram
equalization. In the Histogram equalization, the pixels are
spreading uniformly. Usually Histogram Equalized resultant
images are unnatural look that means it over enhances the
images. The WTHE reduces the effect caused by the
histogram equalization. When the contrast of an image is
enhanced, as an upshot the noise induced in the image will be
high and the image looks unnatural. In this paper, the
proposed technique can be effectively used to improve the
contrast of images and filter out the noises in the image and
preserves the edges. Computed quantitative measures AMBE
and EME value for histogram equalization and weighted
threshold histogram equalization are listed in table I.
I
J
A
E
S
T
R.Sharmila* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 2, 206 - 211
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 42
Comparison of AMBE and EME values show that the
proposed methods outperform well compared to the histogram
equalization. For the proposed method, the computed value of
PSNR and MSE value for WTHE followed by Median and
ISMF are listed in Table II .From the table, we conclude that
the PSNR and MSE values are better than the median filter.

VII CONCLUSION
A novel approach to image enhancement technique using
WTHEISMF is proposed which is better than the existing
histogram equalization method and WTHE followed by
median filtering. The proposed method depicts that the image
enhancement and restoration parameters like AMBE, EME,
PSNR and MSE are outperform well as compared to the
existing method. It effectively reduces over-enhancement,
level clamping artifacts and impulse noise produced due to
enhancement. The proposed method is computationally easy
and can be implemented on FPGA and work well on real time
processing system.

REFERENCES
[1] R.Sherrier and G.Johnson, Regionally adaptive histogram
equalization of the chest, IEEE Tran.Med.Image.,vol.MI-
06,no.1,pp.1-7,Jan. 1987.
[2] A.Polesel, G.Ramponi, and V.Mathews, Image enhancement via
adaptive unsharp masking,IEEE Trans.Image Process., vol.9. no.3,
pp.505-510, Mar.2000.
[3] Tarik Arici, salih Dikbas.A Histogram modification framework
and its application for image contrast enhancement,IEEE Trans.
Image Process., vol.18, no.9,pp. 1921-1935,Sep 2009.
[4] Q. Wang and R. K. Ward, Fast image/video contrast
enhancement based on weighted thresholded histogram equalization,
IEEE Trans.Consum. Electron., vol. 53, no. 2, pp. 757764, May
2007.
[5] P.Jagaatheeswari, S.Suresh kumar and M.Rajaram,A novel
approach for contrast enhancement based on histogram equalization
followed by median filter,APRN.,vol. 4,no. 7, pp.41-45, Sep 2009
[6] S.Zhang,M.A.Karim,A new impulse detector for switching
median filters,IEEE Signal Processing Letters,vol. 9,no.11,pp.360-
363,2002.
[7] Rajoo pandey,An improved switching median filter for
uniformly distributed impulse noise removal, World Academy of
Science. pp.349-351., 2008.
[8] Y-T. Kim, Contrast enhancement using brightness preserving
bihistogram equalization, IEEE Trans. on Consumer Electronics,
vol.43,no. 1pp. 1 8, Feb. 1997.
[9] Y. Wan, Q. Chen, and B-M. Zhang, Image enhancement based
on equal area dualistic sub-image histogram equalization method,
IEEE Trans. on Consumer Electronics, vol. 45,no. 1, pp. 68 75
,Feb. 1999.
[10] S-D Chen and A. Ramli, Minimum mean brightness error bi-
histogram equalization in contrast enhancement, IEEE Trans. on
Consumer Electronics, vol. 49,nNo. 4, pp. 1310 1319 ,November.
2003.
[11] Y-T. Kim, Quantized bi-histogram equalization, ICASSP-97,
vol. 4, pp. 2797 2800, April 1997.
[12] R. C. Gonzalez and R. E.Woods, Digital Image processing, 2nd
ed.Englewood Cliffs, NJ: Prentice-Hall, 2002.
[13] Kim, Y, Contrast Enhancement using Brightness preserving
Bi-histogram equalization, IEEE Transactions on Consumer
Electronics, vol.43, Issue 1, pp 1-8,1997.
[14] G.R.Arce and M.P.McLoughlin,Theoretical analysis of the
MAX/Median filter,IEEE Trans.Acoust.Speeach,Signal
Processing,vol.ASSP-35,pp.60-69,1987.
[15] N. Phanthuna, F. Cheevasuvit and S. Chitwong, Contrast
enhancement for minimum mean brightness error from
histogram partitioning, ASPRS, Mar.2009.

AUTHORS PROFILE
She is graduated B.E (EEE) from Bharathiyar
University Coimbatore in the year 1998, Post
graduated in M.E (VLSI Design) from Anna
University Chennai in the year 2004. Currently she
has been working as Assistant Professor in
Electronics and Communication Engineering, Rajiv
Gandhi College of Engineering and Technology,
Puducherry. She has been teaching VLSI Design, Embedded
Systems, Microprocessor and Microcontrollers for PG and UG
students. She authored books on VLSI Design. She has published
several papers on national conference and symposium. She is the
guest faculty for Pondicherry University for M.Tech Electronics. He
has been actively guiding PG and UG students in the area of VLSI,
Embedded and image processing. She has received the best teacher
award for the year 2006 and 2007. Her research interests are Analog
VLSI Design, Low power VLSI Design, Testing of VLSI Circuits,
Embedded systems and Image processing. She is a member of ISTE.

Sharmila.R received her B.E (Electronics &
Communication Engineering) from Annamalai
University in 2006 & M.E (Computer &
Communication Engineering) from Anna University
in 2010. Currently she has been working as a lecturer
in the Department of Electronics and Communication Engineering in
Rajiv Gandhi College of Engineering & Technology Puducherry .Her
area of interests are image processing and wireless sensor network.
She is a member of ISTE.



I
J
A
E
S
T
R.Sharmila* et al. / (IJAEST) INTERNATIONAL JOURNAL OF ADVANCED ENGINEERING SCIENCES AND TECHNOLOGIES
Vol No. 7, Issue No. 2, 206 - 211
ISSN: 2230-7818 @ 2011 http://www.ijaest.iserp.org. All rights Reserved. Page 43