International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3231
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
A Fast and Robust Hybridized Filter For Image De-Noising
1Ramandeep Kaur
1Student of M.Tech IT, Guru Kahi University,
Talwandi Sabo, India
2Er.Rachna Rajput
2Assistant Professor, CSE, Guru Kashi University,
Talwandi Sabo, India
Abstract - Salt & pepper noise degrades the quality
of the image by hiding the details of objects in the
image and also causes damages to the colour
quality of an image. Noise removal is an important
task in image processing. The noise has to be
removed to obtain the good quality image after
removing the salt and pepper noise. But the
existing salt & pepper noise removal filter like
median, wiener and order static filters are not
capable of reproducing object details in image with
higher accuracy. In this Dissertation work, I have
developed a new hybridized filter for the removal
of salt & pepper noise. This proposed filter will
remove the noise with minimum image quality
degradation. I propose the development of an
advanced salt & pepper noise removal filter using
effective statistic and image processing methods to
remove the noise along with support vector
machine (SVMs) that is effectively do the job by
reproducing the deep image details after removing
the noise, which enhance the quality of image than
the existing filters. The results presented show that
this filter slightly outperforms previous salt and
pepper filters, both in quality and in edge
preservation. To compare results of all existing
filters with new hybridized filter, I use comparison
parameters like PSNR, and MAE.
KEYWORDS– Salt and Pepper Noise, Median
Filter, Order Static Filter, SVM, PSNR, MAE.
I. INTRODUCTION
Images taken with both digital cameras and
conventional film cameras will pick up noise from
a variety of sources. Many further uses of these
images require that the noise will be (partially)
removed – for aesthetic purposes as in artistic work
or marketing, or for practical purposes such as
computer vision. There are various types of noise
which can affect an image such as Salt and Pepper
noise, Gaussian noise, Shot noise etc. In salt and
pepper noise (sparse light and dark disturbances),
pixels in the image are very different in color or
intensity from their surrounding pixels; the defining
characteristic is that the value of a noisy pixel bears
no relation to the color of surrounding pixels [1].
Generally this type of noise will only affect a small
number of image pixels.
When viewed, the image contains dark and white
dots, hence the term salt and pepper noise. Typical
sources include flecks of dust inside the camera and
overheated or faulty CCD(charge-coupled device)
[2].
Image processing is the study of any algorithm that
takes an image as input and returns an image as
output. Image processing is any form of signal
processing for which the input is an image, such as
a photograph or video frame the output of image
processing may be either an image or a set of
characteristics or parameters related to the image.
The digital image is processed by a computer to
achieve the desired result. Image enhancement
improves the quality (clarity) of images for human
viewing. Removing blurring and noise, increasing
contrast, and revealing details are examples of
enhancement operations. For example, an image
might be taken of an endothelial cell, which might
be of low contrast and somewhat blurred. Reducing
the noise and blurring and increasing the contrast
range could enhance the image. The original image
might have areas of very high and very low
intensity, which mask details. An adaptive
enhancement algorithm reveals these details.
Adaptive algorithms adjust their operation based on
the image information (pixels) being processed. In
this case the mean intensity, contrast, and sharpness
(amount of blur removal) could be adjusted based
on the pixel-intensity statistics in various areas of
the image.
An image may be described as a two-dimensional
function
I=f(x, y)
Where x and y are spatial coordinates. Amplitude
of f at any pair of coordinates (x, y) is called
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3232
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
intensity I or gray value of the image. When spatial
coordinates and amplitude values are all finite,
discrete quantities, the image is called digital
image. Digital image processing may be classified
into various sub branches based on methods whose:
• Inputs and outputs are images.
• Inputs may be images where as outputs are
attributes extracted from those images.
Digital images are form of visual information
captured or transmitted using camera or other
imaging system. The received image might be
corrupted due to the presence of noise. Image noise
reduction without structure degradation is perhaps
the most important low-level image processing
task. Faulty sensors, optic imperfectness,
electronics interference, and data transmission
errors may introduce noise to digital images. In
considering the signal-to-noise ratio over practical
communication media, such as microwave or
satellite links, there can be degradation in quality
due to low received signal power. Based on
trichromatic color theory, color pixels are encoded
as three scalar values, namely, red, green and blue
(RGB color space). Since each individual channel
of a color image can be considered as a
monochrome image, traditional nonlinear image
filtering techniques often involved the application
of scalar filters on each channel separately.
However, this disrupts the correlation that exists
between the color components of natural images.
As such the color noise model should be considered
as a 3-channel perturbation vector in color space.
Image noise is the random variation of brightness
or colours information in images produced by the
sensor and circuitry of a scanner or digital camera.
Image noise can also originate in film grain and in
the unavoidable shot noise of an ideal photon
detector. Although these unwanted Fluctuations
became known as "noise" by analogy with
unwanted sound they are inaudible and such as
dithering.
The types of Noise are following:-
• Amplifier noise (Gaussian noise)
• Salt-and-pepper noise
• Speckle noise etc.
Salt and pepper noise also called as an impulse
noise. It is also referred to as intensity spikes.
Mainly while transmitting data we will get this salt
and pepper noise. It has only two possible values, 0
and 1. The probability of each value is typically
less than 0.1. The corrupted pixel values are set
alternatively to the maximum or to the minimum
value, giving the image a “salt and pepper” like
appearance as salt looks like [13]white(one) and
pepper looks as black(zero) for binary ones. Pixels
which are not affected by noise remain unchanged.
For an 8-bit image, the typical value for pepper
noise is 0(minimum) and for salt noise
255(maximum). This noise is generally caused in
digitization process during timing errors,
malfunctioning of pixel elements in the camera
sensors, faulty memory locations.
SUPPORT VECTOR MACHINE
Support Vector Machine (SVM) was first heard in
1992, introduced by Boser, Guyon, and Vapnik in
COLT-92. Support vector machines (SVMs) are a
set of related supervised learning methods used for
classification and regression [5]. They belong to a
family of generalized linear classifiers. In another
terms, Support Vector Machine (SVM) is a
classification and regression prediction tool that
uses machine learning theory to maximize
predictive accuracy while automatically avoiding
over-fit to the data. Support Vector machines can
be defined as systems which use hypothesis space
of a linear functions in a high dimensional feature
space, trained with a learning algorithm from
optimization theory that implements a learning bias
derived from statistical learning theory. Support
vector machine was initially popular with the NIPS
community and now is an active part of the
machine learning research around the world. SVM
becomes famous when, using pixel maps as input;
it gives accuracy comparable to sophisticated
neural networks with elaborated features in a
handwriting recognition task [9]. It is also being
used for many applications, such as hand writing
analysis, face analysis and so forth, especially for
pattern classification and regression based
applications. The foundations of Support Vector
Machines (SVM) have been developed by Vapnik
[27] and gained popularity due to many promising
features such as better empirical performance. The
formulation uses the Structural Risk Minimization
(SRM) principle, which has been shown to be
superior, to traditional Empirical Risk
Minimization (ERM) principle, used by
conventional neural networks. SRM minimizes an
upper bound on the expected risk, where as ERM
minimizes the error on the training data. It is this
difference which equips SVM with a greater ability
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3233
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
to generalize, which is the goal in statistical
learning. SVMs were developed to solve the
classification problem, but recently they have been
extended to solve regression problems [5].
II. Median filter
A digital filter is a system that performs
mathematical operations on a sampled. There are
two types of filters that are used to remove
different type of noises from digital images.
Linear filters are used to remove certain type of
noise. Gaussian or Averaging filters are suitable for
this purpose. These filters also tend to blur the
sharp edges, destroy the lines and other fine details
of image, and perform badly in the presence of
signal dependent noise. Non-linear filters have
quite different behaviour compare to linear filters.
For non-linear filters, the filter output or response
of the filter does not obey the principals of scaling
and shift invariance. In this project I use Average
filter, and Median filter. Average filter is linear
filters and a median filter is a non-linear filter.
Image de-noising is a vital image processing task
i.e. as a process itself as well as a component in
other processes. There are many ways to de-noise
an image or a set of data and A method exists. The
important property of a good image de-noising
model is that it should completely remove noise as
far as possible as well as preserve edges.
Traditionally, there are two types of models i.e.
linear model and non-liner model. The benefits of
linear noise removing models is the speed and the
limitations of the linear Models are the models are
not able to preserve edges of the images in a
efficient manner i.e the edges, which are
recognized as discontinuities in the image, are
smeared out. On the other Hand, Non-linear models
can handle edges in a much better way than linear
models.
Figure: 1 Types of Filter
A median filter comes under the class of nonlinear
filter. The best known order statistics filter is the
median filter that replaces the value of a pixel by
the median of their neighbourhood pixels.Median
Filters are very effective in removing impulse noise
at low density levels.
The median filter follows the moving window
principle for filtering. A 3 × 3, 5 × 5 or 7 × 7 kernel
of pixels is scanned over pixel matrix of the
complete image. The median of the pixel values
within the window is computed, and therefore the
center pixel of the window is replaced with the
computed median. Median filtering is completed
by, initial sorting all the pixel values from the
surrounding neighbourhood into numerical order so
substitution the pixel being considered with the
centre pixel value. Note that the median value must
be written to a separate array or buffer in order that
the results are not corrupted because the method is
performed.
The below process illustrates the methodology of
median filtering. 5x5 mask and the pixel values of
image in the neighbourhood of considered noisy
pixel are
1
2
5
1
4
7
1
7
5
1
1
1
1
5
0
1
2
0
1
1
5
1
5
0
1
0
8
1
1
8
1
2
2
1
3
2
1
4
0
1
0
7
1
1
2
1
1
2
1
5
2
1
2
8
1
3
4
1
1
2
1
3
4
1
5
5
1
5
5
1
9
8
1
4
5
Table: 1 Median Values in the Neighbourhood Of
140
Algorithm for Median Filter
1. Step 1. Select a two dimensional window W of
size 3*3. Assume hat the pixel being processed
is Cx,y.
2. Step 2. Compute Wmed the median of the
pixel values in window W.
3. Step 3. Replace Cx,y by Wmed.
4. Step 4. Repeat steps 1 to 3 until all the pixels
in the entire image are processed.
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3234
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
Advantage:
a. It is easy to implement.
b. Used for de-noising different types of noises. [5]
Disadvantage:
a.Median Filter tends to remove image details
while reducing noise such as thin lines and corners.
b. Median filtering performance is not satisfactory
in case of signal dependant noise. To remove these
difficulties different variations of median filters
have been developed for the better results.
III. PARAMETERS ANALYZED
In order to determine the performance of various
noise removal algorithms the following parameters
are analyzed:
A.Peak Signal-to-Noise Ratio
B.Mean Absolute Error
A. Peak Signal-To-Noise Ratio
Peak signal-to-noise ratio, often abbreviated 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 most commonly used as a measure of
quality of reconstruction of lossy compression
codecs (e.g., for image compression). The signal in
this case is the original data, and the noise is the
error introduced by compression. When comparing
compression codecs it is used as an approximation
to human perception of reconstruction quality,
therefore in some cases one reconstruction may
appear to be closer to the original than another,
even though it has a lower PSNR (a higher PSNR
would normally indicate that the reconstruction is
of higher quality) [6].
where MAX is the maximum possible pixel value
of the image.
b. Mean Absolute Error
It is a quantity used to measure how close forecasts
or predictions are to the eventual outcomes. The
mean absolute error is given by
As the name suggests, the mean absolute error is an
average of the absolute error
,
Where are the prediction and the true value.
The mean absolute error is a common measure
of forecast error in time series analysis, where the
terms "mean absolute deviation" is sometimes used
in confusion with the more standard definition
of mean absolute deviation
IV. PROPOSED MODEL
In this research, we will work on the development
of a new method for the removal of salt & pepper
noise by creating a new hybridized filter using
existing and/or new noise removal filters. The
proposed filter will remove the noise with no or
minimum image quality degradation. Salt & pepper
noise degrades the quality of the image by hiding
the details of objects in the image and also causes
damages to the colour quality of an image. The
noise has to be removed to obtain the good quality
image after removing the salt and pepper noise. But
the existing salt & pepper noise removal filter like
median filter are not capable of reproducing object
details in image with higher accuracy. We propose
the development of an advanced salt & pepper
noise removal filter using effective statistic and
image processing methods to remove the noise
along with support vector machine (SVMs) that
will effectively do the job by reproducing the deep
image details after removing the noise, which will
enhance the quality of image than the existing
filters.
Algorithm for Hybridized Filter
1. Img Load Image ()
2. FiltSvm Load SVM Filter
3. ImageRestored FiltSvm(Img)
4. Allocate outputPixels ()
a. If not lastSegment(feof(blockSize))
b. clrArr ColorArray()
c. abrIdx abnormIndex(clrArr)
d. nIdx neuralize(clrArr,abrIdx)
5. sortClrArr sort(nIdx)
6. outputPixel level(sortClrArr)
This proposed method will help pathologists
identify the characteristic of pyloric stenos is. For
this reason, the paper is separately as follows. First,
as mention in introduction, the brief review about
medical image is guided and summary about filters
technique. Second is methodology of doing the
tasks. Third, the three example tests based on these
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3235
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
techniques are tried. Meanwhile, we also made the
comparisons with median filter, wiener filter, and
order-static filter. Finally, some conclusions are
made and discussed.
Hybridized De-Noising filter Flow Chart
1: Flow chart for Hybridized De-Noising filter
In results we are taking an original coloured image
and then convert it into grey scale image that is
black and white image. Then we are adding salt and
pepper noise into the original image.
Figure 1: Showing Original Image and its planes
In figure 1 we are taking an original Colour image.
Image name is tomato. Then we show its three
RGB planes these planes are red, green and
blue.RGB means Red, Green and Blue planes.
Figure 2: Images with salt and pepper noise on
different planes
Now we are adding salt and pepper noise into an
original coloured image. Then we display results on
the image RGB (Red, Green, Blue) planes. salt and
pepper noise is present in the image in the form of
black and white dots. Which corrupt the image and
hide the details of an image.
Image Acquisition
Load SVM Image Filter
Apply SVM Image Filter
Allocate output pixels and
return image
Extract color array
Extract Abnormal Index
Neuralize the color array on
the basis of abnormal index
Sort and level the image
matrix and return de-noised
image
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3236
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
Figure 3: Results of Median filtering
In this figure we are showing results of median
which is used for noise reduction on different
planes of an image.
Median filter is a spatial filtering operation, so it
uses a 2-D mask that is applied to each pixel in the
input image. It is used to remove defects and noise
from pictures.
Median filter is much less sensitive than the mean
to extreme values (called outliers).
Figure 4: Results of Order-static filtering
In this figure we are showing results of order static
filter which is used for noise reduction on different
planes of an image. The classical order-static
masking is the method that using nonlinear Filter.
The order-static masking method operates by
adding a fraction of the high pass filtrated version
of the input image to the original one. This operator
is sensitive to noise due to the presence of the
linear high pass filter which cannot discriminate
signal from noise. Moreover it perceptually
enhances image more in dark areas than in lighter
ones.
Figure 5: Result of Hybridized Filter
This is new filter named as hybridized filter used to
de-noising the image and its different planes.
Figure 6: Results SVM
In figure 5 we are using SVM (Support Vector
Machine) for noise estimation into the image.
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3237
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
Figure 7: Comparisons of all filters
In figure 7 Results of the various image de-noising
filters have been shown. The images are added up
with noise during the acquisition or transmission
processes. In order to remove the noise from the
images, the image matrix has to be normalized. The
normalization of the images is also called as image
de-noising. The filters used in this research are
order static, median and the Hybridized filter. The
results have proved the effectiveness of the
Hybridized filter when compared to the other
image de-noising filters.
VII. COMPARISONS OF ALL FILTERS
We are comparing the results of filters with MAE
AND PSNR Comparison Parameters.
For image 1:
Table 1: Comparison of Filters for image1
Filter Name MAE PSNR
Median Filter: 1.269264 78.934946
Order Static
Filter
22.618818 75.480454
hybridized
Filter
-0.222963 82.561588
For image 2
Table 2: Comparison of Filters for image 2
Filter Name MAE PSNR
Median Filter: 4.738779 94.629591
Order Static
Filter
37.964697 91.175099
hybridized
Filter
4.119984 98.256233
These tables are shows the comparison of all filters
with comparison parameters PSNR and MAE
values for image1. By these values we get better
results with Hybridized filter then Median and
Order static filter. Hybridized filter is smoother and
shows the details of image than the other ones.
VIII. CONCLUSION AND FUTURE SCOPE
1 CONCLUSION
Salt & pepper noise degrades the quality of the
image by hiding the details of objects in the image
and also causes damages to the colour quality of an
image. The noise has to be removed to obtain the
good quality image after removing the salt and
pepper noise. But the existing salt & pepper noise
removal filter like median filter, order static are not
capable of reproducing object details in image with
higher accuracy. In this dissertation, I create a new
filter that is hybridized filter using existing for the
removal of salt & pepper noise. The proposed filter
will remove the noise with no or minimum image
quality degradation. We propose the development
of an advanced salt & pepper noise removal filter
using effective statistic and image processing
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3238
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
methods to remove the noise with hybridized filter
which will enhance the quality of image than the
existing filters. Quality of image is also be
improved and values of PSNR and MAE are also
give better results as compare to the values of
existing filters.
2 FUTURE SCOPE
In the future, various techniques can be considered
to incorporate in this scheme to further improve the
performance and preserve more edges in both
highly and lowly corrupted images. We also can
develop a filter that can completely remove high
density noise from an image and work on the
details of an image. In future any one can improve
the performance of de-noising filter and also show
improvement in Comparison parameters values like
PSNR, MSE, and MAE. To improve de-noising
along the edges as the method we used did not
perform so well along the edges. In future salt and
pepper noise is also removing from audio and video
signals.
REFERENCES
[1] Amandeep Kaur, Karamjeet Singh (2010),
“Speckle Noise Reduction By Using Wavelets”,
NCCI 2010 -National Conference On
Computational Instrumentation Chandigarh, India.
[2] Abhishek Kesharwani Sumit Agrawal Mukesh
Kumar Dhariwal (2013), “An Improved Decision
based Asymmetric Trimmed Median Filter for
Removal of High Density Salt and Pepper Noise”,
International Journal of Computer Applications
(0975 – 8887) Volume 84 – No 8.
[3] Chandra Sekhar Panda, and Prof. (Dr.)
SrikantaPatnaik (2010), “Filtering Corrupted
Image and Edge Detection in Restored Grayscale
Image Using Derivative Filters”, International
Journal of Information Processing, Volume (3):
Issue (3).
[4] Chanchal Srivastava, Saurabh Kumar And O.P.
Singh (2013), “performance comparison of various
filters and wavelet transform for image de-noising”
IOSR Journal of computer Engineering (IOSR-
JCE), p-ISSN: 2278-8727Vol. 10,Issue. 1.
[5] Er. Nancy, Er. Sumandeep Kaur (2013),
“Comparative Analysis and Implementation of
Image Enhancement Techniques Using MATLAB”
International Journal of Computer Science and
Mobile Computing, Vol. 2, Issue.4.
[6] Irphan Ali Shaik, 2Mirza shafi shahsavar,
3K.J.Silva Lorraine, 4Ajesh kumar vishwanadham
(2012), “Impulse Noise Detection and Filtering
Based on Adaptive Weighted Median Filter” ISSN:
2278-4721, Vol. 1, Issue 8.
[7] J.Harikiran, B.Saichandana, and B.Divakar
(2010), “Impulse Noise Removal in Digital
Images”, International Journal of Communication
Systems, Volume 10– No.8, 39-42.
[8] Mrs. C. Mythili and Dr. V. Kavitha (2011),
“Efficient Technique for Color Image Noise
Reduction”, TRBJ ACM - ISWSA; ISSN: 2078-
7952 (print); 2078-7960 (online).
[9] Mr. Salem Saleh Al-amri, Dr. N.V.
Kalyankarand Dr.Khamitkar S.D (2010), “A
Comparative Study of Removal Noise from
Remote Sensing Image” IJCSI International
Journal of Computer Science Issues, Vol. 7, Issue.
1, No. 1.
[10] Om Prakash Verma and Shweta Singh
(2013),“A Fuzzy Impulse Noise Filter Based on
Boundary Discriminative Noise Detection”,
Journal of Information Process System, Vol.9,
No.1.
[11] Pawan Patidar and Manoj Gupta (2010),
“Image De-noising by Various Filters for Different
Noise”, International Journal of Computer
Applications (0975 – 8887) Volume 9– No.4,
November.
[12] P.E.Ng and K.K.Ma(2006),‟ Aswitching
median filter with boundary discriminative noise
detection for exteremely corrupted image,”IEEE
Trans. Image process, vol.15,no.6,PP.1506-1516.
[13] Priyanka Kamboj and Versha Rani (2013), “A
Brief Study Of Various Noise Model And Filtering
Techniques”, Journal of Global Research in
Computer Science, Volume 4, No. 4.
[14] Ravinder Singh Balwinder Singh and
Harmandeep Singh (2013), “Removal of High
Density Salt & Pepper Noise in Noisy color Images
using Proposed Median Filter” International
Journal of Advanced Research in Computer
Science and Electronics Engineering (IJARCSEE)
Volume 2, Issue 2.
[15] Sanjeev Kumar, Asutosh Kar and Mahesh
Chandra (2014), “SVM Based Adaptive Median
Filter Design for Face Detection in Noisy Images”,
International Conference on Signal Processing and
Integrated Networks (SPIN).
[16] S.Gopi Krishna, T. Sreenivasulu Reddy
andG.K.Rajini, “Removal of High Density Salt and
Pepper Noise Through Modified Decision Based
Unsymmetric Trimmed Median Filter”
International Journal of Engineering Research and
International Journal of Science, Engineering and Technology Research (IJSETR), Volume 3, Issue 12, December 2014
3239
ISSN: 2278 – 7798 All Rights Reserved © 2014 IJSETR
Applications (IJERA) Vol. 2, Issue 1, Jan-Feb
2012.
[17] Srinivasan, K.S., Ebenezer, D. (2007), “A new
fast and efficient decision based algorithm for
removal of high-density impulse noises”, IEEE
Signal Process.Lett.14(3), Pp.189–192.
[18] Sucharita Padhi Debananda Padhi, and Bratat
iMohanty (2012), “Order Statistic Filters for
removing Salt-and-Pepper Noise of Images: A
Comparative Study” International Journal of
Advanced Research in Computer Science and
Software Engineering Volume 2, Issue 1.
[19] Sukomal Mehta, Sanjeev Dhull (2012), “fuzzy
based median filter for grey-scale images”,
International journal of engineering science &
advanced technology, ISSN:2250-3676,Volume-
2,Issue-4.
1.Author Profile
Ramandeep Kaur received the M.SC(IT) Master
Degrees in Information Technology degree from
T.P.D punjabi university nebhourhood campus
rampura-phul in 2012.Currently she is pursuing
m.tech degree in information technology from
Guru Kashi University, Talwandi Sabo, Bathinda
(Punjab). Her research interests include digital
image processing and Image Enhancement.