IRJET- Image Restoration using Adaptive Median Filtering

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 10 | Oct 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 841
Image Restoration using Adaptive Median Filtering
Hetvi Soni[1], Darshana Sankhe[2]
Student[1], Professor[2], Dept. Of Electronics Engineering, D.J Sanghvi College of Engineering, Mumbai, India
----------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – In the field of image processing, image
restoration is the most essential task. Image often gets
corrupted due to which there is presence of noise in the
image. Generally median filter is used to remove the
presence of such noise but median filter works fine for
about intensity of 20% noise in the image. So in order
to get a better image restoration we can use another
image restoration technique which is adaptive median
filtering which works very well for noise intensity
beyond 20% .The benefit of adaptive filter over median
filter is that it does not erode away edges or small
details in the image.
Key Words: Adaptive median filter, Image
Processing, Median filter, Salt and Pepper Noise,
Gaussian Noise PSNR.
1. INTRODUCTION
In the field of image processing image gets corrupted
during image transmission or acquisition stage due
to various influencing parameters such as faulty
device etc. Such type of corruption gives rise to noise
in the image which affects the information present in
the image. The various type of noise that arises in the
image is impulse noise also know as salt and pepper
noise, Gaussian noise, and speckle noise to name a
few [1].
In order to retrieve the original information from the
image and eliminate noise from the image we need to
apply certain de-noising techniques. These de-
noising techniques make use of kernel which is made
to convolve over the image and the result obtain
through this convolution is a noise free image. The
size of kernel (window) used for the same varies and
intended output with same also varies.
One such commonly used technique is median
filtering. Median filtering works fine when the noise
intensity is less but it starts to fail when the noise
intensity in the image is high [3].
In order to overcome this problem we can make of
spatial filtering technique. One such Filtering
algorithm used is adaptive median filter. Adaptive
median filtering is better than median as it is a two
step filtering technique. The main advantage of
adaptive median filter is that the behavior of the
adaptive filter changes depending on the
characteristics of the image under filter. Other main
feature of adaptive filter is that it works well not only
for impulse noise but also for speckle noise and
Gaussian noise [3].
2. MEDIAN FILTER
Median filter is the most commonly used filter. It
is a non linear method of filtering. The size of the
kernel can be of nxn size which is made to convolve
or slide over a mxm corrupted image. While
performing this operation the median value of nxn
kernel on the image is obtained and then the value of
a particular pixel is replaced with the median value
of the nxn kernel.
Fig -1: Sorting in median filter.
2.1. Drawbacks of median filter:
 Effective only when the noise is impulse
noise (salt and pepper).
 Its output quality deteriorates when the
noise is more than 20%.
 It does not work efficiently when the
spatial density of noise is high.
 For large kernel size, there is no proper
smoothening of the image instead
valuable information from the image
gets blur.

2. ADAPTIVE MEDIAN FILTER
The adaptive median filter is designed to
eliminate the drawbacks faced by the standard
median. The main advantage of adaptive median
filter is the size of the kernel surrounding the
corrupted image is variable due to which better
output result is obtained.
The other main advantage of adaptive filter is that
unlike median filter it does not replace all the pixel
values with the median value.
The working of adaptive filter is a two step process;
in the first step its finds the median value for the
kernel and in the second step it checks whether the
current pixel value is a impulse (salt and pepper
noise) or not. If pixel value is corrupted then it
changes it value with median or else it retains the
value of the gray scale pixel.
By doing this, one can ensure that only the pixels
with impulse noise is changed while all other pixel
values are retained as they are.
3.1. Implementation of adaptive median filter
Zmin = Minimum gray level value in Sxy.
Zmax = Maximum gray level value in Sxy
Zmed = Median of gray levels in Sxy
Zxy = gray level at coordinates (x, y)
Smax = Maximum allowed size of Sxy
The adaptive median filter works in two levels
denoted Level A and Level B as follows:
Level A: A1= Zmed - Zmin
A2= Zmed - Zmax
If A1 > 0 AND A2 < 0, Go to level B
Else increase the window size
If window size <=Smax repeat level A
Else output Zxy.
Level B: B1 = Zxy – Zmin
B2 = Zxy – Zmin
If B1 > 0 And B2 < 0 output Zxy
Else output Zmed.
3.2. Flowchart:
Fig-2: Algorithm used for adaptive median filter
4. RESULTS:
The standard median filter and adaptive median
is been implemented using MATLAB Software. The
PSNR (peak signal to noise ratio) is been calculated
to understand the efficiency of both the filters.
Basically when more is the PSNR value better is noise
reduction by the filter.
From the calculated values we can observe that the
adaptive median filter has better noise filtering
capacity than the standard median filter.
Table-1: PSNR Value for salt &pepper noise
Noise
percentage
Median
Filter
(3x3)
Median
Filter
(5x5)
Adaptive
Median
Filter
10% 14.9433 14.625 41.8677
20% 12.009 11.8033 41.8702

30% 10.2944 10.1596 41.8816
40% 9.1065 8.9761 41.8675
50% 8.0873 8.0011 41.8810
Table-2: PSNR Value for Gaussian noise
Noise
percentage
Median
Filter
(3x3)
Median
Filter
(5x5)
Adaptive
Median
Filter
10% 19.5683 18.6263 43.2747
20% 19.7813 18.8621 44.4432
30% 20.5014 19.5324 45.4093
40% 21.8413 20.7830 46.0971
50% 23.3442 22.1562 46.5204
Table-3: Output for salt & pepper noise.
Table-4: Output for Gaussian noise.
CONCLUSIONS
After implementing both standard and adaptive
median filter, we can observe that the adaptive
median filter works accurately for the impulse noise
that is salt and pepper noise.
For Gaussian noise the adaptive median filter gives
much better output than the standard median but
we can also observe that the output results are not so
satisfactory and accurate ,so there needs to be some
more modification to be done in the proposed model
to obtained better results for Gaussian noise.
Original image
Median 3x3
noisy image
Median 5x5
Adaptive median filter
Original image noisy image
median 3x3 Median 5x5
Adaptive median filter

REFERENCES
[1] Rafael C. Gonzalez and Richard E. Woods
Digital Image Processing, 2001, pp.220-225.
[2] R. Boyle and R. Thomas Computer Vision: A
First Course, Blackwell Scientific Publications,
1988, pp. 32 - 34.
[3] Suman Shrestha, Image denoising using new
adaptive based median filter ,Signal & Image
Processing : An International Journal (SIPIJ)
Vol.5, No.4, August 2014
[4] Uk, Ijeacs. (2017). Performance Assessment
of Several Filters for Removing Salt and
Pepper Noise, Gaussian Noise, Rayleigh Noise
and Uniform Noise. International Journal of
Engineering and Applied Computer Science
(IJEACS). 02. 176-180.
10.24032/ijeacs/0206/01.
[5] Afrose, Zinat. (2012). Relaxed Median Filter:
A Better Noise Removal Filter for Compound
Images. International Journal on Computer
Science and Engineering (IJCSE).
[6] Ammayappan, Sathesh. (2018). A Nonlinear
Adaptive Median Filtering Based Noise
Removal Algorithm.

IRJET- Image Restoration using Adaptive Median Filtering

More Related Content

What's hot (20)

Similar to IRJET- Image Restoration using Adaptive Median Filtering (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Image Restoration using Adaptive Median Filtering