A Modified Weighted Based Filter for Removal of Random Impulse Noise (MWB)
J.K. Mandal
Department of Computer Science & Engineering
University of Kalyani
Kalyani, Nadia, West Bengal
e-mail: jkm.cse@gmail.com
Aparna Sarkar
Department of Computer Science & Engineering
University of Kalyani
Kalyani, Nadia, West Bengal
e-mail:sarkaraparna09@gmail.com.
Abstract— In this paper an effort has been made to device an
algorithm for highly corrupted images. In this paper a novel
Modified Weighted Based(MWB) filter has been proposed,
which is based on the weighted differences between the
current pixel and its neighbors aligned with four main
directions. Simulations showed that the MWB filter provides
optimal performances of suppressing impulse with high
noise level which may enhance the performance in terms of
removal of random-valued impulse noise compared to the
directional weighted median (DWM) filter [1] along with
other filters like signal-dependent rank order mean (SD-
ROM) filter[6], multistate median (MSM) filter[2], adaptive
center weighted median (ACWM) filter[8], pixel-wise MAD
(PWMAD) filter[7] etc.
Keywords- Modified Weighted Based Filter (MWBF);
PSNR; SNR; Median; DWM
I. INTRODUCTION
Images are often corrupted by impulse noises during
acquisition and transmission [5, 6]. Based on the noise
values, the noise can be classified as the easier-to-restore
salt-and-pepper noise and the more difficult random-
valued impulse noise [2]. Among all the methods for
removal of impulse noise, the median filter [3, 4] is used
widely because of its effective noise suppression
capability and high computational efficiency [7].
However, it uniformly replaces the gray-level value of
every pixel by the median of its neighbors. Consequently,
some desirable details are also removed, especially when
the window size is large. In order to improve the median
filter, many filters with an impulse detector are proposed,
DWM filter [1] is one of them. DWM filter performs
much better than the other median-based filters in
removing random-valued impulse noise, especially when
the noise level is as high as 60%.
In this paper a filter is used for removal of random-
valued impulse noise for which the performance may be
comparable/better with DWM filter. This filter has been
proposed to obtain optimal performance for highly
corrupted images.
The organization of this paper is as follows. The new
impulse detector is formulated in section I. Section II
described the filtering framework. Section III provides the
used algorithm. Section IV provides a number of
experimental results to demonstrate the performance of the
proposed MWB filter. Conclusions are drawn in section V
and that of references in section VI.
II. IMPULSE DETECTOR
Fig. 1 shows the edges aligned with four main
directions. The edges separate the smoothly varying areas
of a noise free image.
Let S
k
(k=1 to 4) denote a set of coordinates aligned
with the k
th
direction centered at (0,0), i.e.,
S
1
={(-2,-2), (-1,-1), (0,0), (1,1), (2,2)}
S
2
={(0,-2), (0,-1), (0,0), (0,1), (0,2)}
S
3
={(2,-2), (1,-1), (0,0), (-1,1), (-2,2)}
S
4
={(-2,0), (-1,0), (0,0), (1,0), (2,0)} (1)
Now in a 5×5 window centered at (i, j), for each
direction, we define d
i, j
(k)
as the sum of all absolute
differences of gray-level values between y
i+s, j+t
and y
i, j
with (s,t) є S
k
0
, where S
k
0
= S
k
\ (0,0) for all k from 1 to 4.
The gray-level values of any two pixels should be close, if
the spatial distance between them is small. We will weight
the absolute differences between the two closest pixels
with a larger value w
s,t
is very large, it will cause that
d
i,j
(k)
is mainly decided by the differences corresponding
to w
s,t
. Thus we have eq.2.
Fig. 1: Alignment of edges in four directions
2011 Second International Conference on Emerging Applications of Information Technology
978-0-7695-4329-1/11 $26.00 © 2011 IEEE
DOI 10.1109/EAIT.2011.77
173