International Journal of Engineering Research & Technology (IJERT)

ISSN: 2278-0181
Vol. 2 Issue 4, April - 2013

Comparison Of Different Techniques Of Digital Image Denoising

Pooja Sharma1, Amandeep Kaur2

1
Student M.Tech ECE Punjabi University Patiala
2
Assistant Professor Department of ECE Punjabi University Patiala

Abstract an image [1]-[2]. In denoising there is always a trade-

off between noise suppression and preserving actual
Digital images play very important role in daily life image discontinuities. To remove noise without
applications as well as in the areas of research and excessive smoothing of important details, a denoising
technology .But main disadvantage of digital images technique needs to be spatially adaptive. Different
is that noise is added during their transmission. techniques are used depending on the noise model.
Noise in the digital images can be removed by Due to sparsity, an edge detection and
several techniques. Selecting appropriate techniques multiresolution property the wavelet representation
for de-noising plays an important role. Several naturally facilitates such spatially adaptive noise
techniques are proposed for image de-noising and filtering [3].
each technique has its advantages and
RRTT
disadvantages. Mainly two types of denoising
2. Noise models
techniques are there .One is spatial domain filtering
and other is transform domain filtering. Choice of
IIJJEE

denoising technique is dependent upon the type of Different type of noise is present in digital
noise which is present in the digital image. In this images .Noise may be present in additive or
paper after a brief introduction, overview of different multiplicative form [4].
denoising techniques and their analysis is given.
2.1. Additive Noise Model

1. Introduction Noise signal that is additive in nature gets

added to the original signal to produce a corrupted
In digital imaging, the acquisition techniques and noisy signal and follows the following model:
systems introduce various types of noises and w(x, y) = s(x,y) + n(x,y) (1)
artifacts. Image denoising is more significant than
any other tasks in image processing, analysis and
applications. Besides the noisy image produces
undesirable visual quality, it also lowers the visibility 2.2. Multiplicative Noise Model
of low contrast objects. Hence noise removal is
essential in digital imaging applications in order to In this model, noise signal gets multiplied to the
enhance and recover fine details that are hidden in the original signal. The multiplicative noise model
data. In many occasions, noise in digital images is follows the following rule:
found to be additive in nature with uniform power in w(x, y) = s(x,y) × n(x,y) (2)
the whole bandwidth and with Gaussian probability where, s(x,y) is the original image intensity and
distribution. Such a noise is referred to as Additive n(x,y) denotes the noise introduced to produce the
White Gaussian Noise (AWGN). It is difficult to corrupted signal w(x,y) at (x,y) pixel location[4].
Suppress AWGN since it corrupts almost all pixels in

www.ijert.org 417
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 2 Issue 4, April - 2013

3. Various type of noise a. Smoothing (linear) spatial filter

Smoothing filters are used for blurring and noise

There are different types of noise is present and
reduction. Blurring may be implemented in
each type of noise has different characteristics.
preprocessing tasks to remove small details from an
image prior to large object extraction. The output of a
3.1. Gaussian noise
smoothing (averaging or lowpass) linear spatial filter
is the average of the pixels contained in the
Gaussian noise is evenly distributed over the
neighborhood of the filter mask. By replacing the
signal. Each pixel in noisy image is the sum of true
value of every pixel in an image by the average of the
pixel value and a random gaussian distributed noise
intensity levels in the neighborhood defined by a
value [4].
filter mask, the resulting image will have reduced
“sharp” transitions in intensities. However, edges
3.2. Salt and Pepper Noise (characterized by sharp intensity transitions) will be
blurred [6].
Salt and Pepper is an impulse type of noise and
is also referred to as intensity spikes. It is generally
b. Order-statistic(nonlinear) filters
caused due to errors in transmission. It has only two
possible values, a high value and a low value. The
Order-statistic filter are nonlinear spatial filters
probability of each is typically less than 0.1. The
whose response is based on ordering (Ranking) the
corrupted pixels are set alternatively to minimum or
pixels in the neighborhood and then replacing the
to maximum values giving the image Salt and
value of the center pixel by the value determined by
Pepper‖ like appearance. The unaffected pixels
RRTT
the ranking result. The median filters are quite
remain unchanged [4].
effective against the impulse noise (salt-and-pepper
noise). The median of a set of values is such that half
IIJJEE

3.3. Speckle Noise the values in the set are greater than the median and
half is lesser than it[6].
Speckle noise [14]-[15] is multiplicative noise.
This type of noise occurs in almost all coherent c. Bilateral Filter (BF)
systems such as SAR images, Ultrasound images etc.
The source of this noise is random interference Recently most popular denoising method is the
between the coherent returns [4]. bilateral filter [7]. The bilateral filter is a nonlinear
weighted averaging filter and also the weights depend
on both the spatial distance and the intensity distance
4. Different denoising techniques with respect to the center pixel. The main feature of
the bilateral filter is its ability to preserve edges while
doing spatial smoothing. The bilateral filter is a
There are basically two approaches for image
robust filter because of its range weight, pixels with
denoising.One is spatial filtering method and other is
different intensities. It averages local small details
transform domain filtering method.
and ignores outliers[1].
The main drawback in bilateral filter is its
4.1. Spatial domain filtering incapacity in eliminating salt-and-pepper noise. The
second drawback of the bilateral filter is that it
This is traditional way to remove noise from produces staircase effect and it is also single
digital images. Spatial domain filtering is further resolution in nature[8].
classified into linear filters and non-linear filters [5].

www.ijert.org 418
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 2 Issue 4, April - 2013

d. Trilateral filter(TF) function φ and a mother wavelet ψ. The finite scale

multiresolution representation of a discrete function
Trilateral filtering is one of the most efficient can be called as a discrete wavelet transforms The
technique for denoisation of the multi-dimensional modeling of the wavelet coefficients can either be
signals in computer graphics. By means of non- deterministic or statistical[16]. Wavelet approach for
linear combination of information of adjacent pixel it noise removal has been successfully exploited by
can smoothen the edges and keeps the details of the several in the past few decades. It has been proved
images. Moreover the process is less iterative and is that the use of wavelets successfully removes noise
applicable for N dimensional images as well. while preserving the signal characteristics, regardless
Trilateral filtering apart from cumulating geometric of its frequency content[12]-[13].
and photometric similarities also implements a new Wavelet transform is a mathematical function
concept of local structural similarities specially in the that analyzes the data according to scale or
inhomogeneous portions. Homogeneous portions are resolution. Noise reduction using wavelets is
however low-pass-filtered out directly without any performed by first decomposing the noisy image into
significant change in composition. Trilateral filtering wavelet coefficients i.e. approximation and detail
shows high performance in high contrast image coefficients. Then, by selecting a proper thresholding
display and edge smoothening. It can extract the value the detail coefficients are modified based on
original information from a noisy image and can the thresholding function. Finally, the reconstructed
predict gradient discontinuities in spacial domain. It image is obtained by applying the inverse wavelet
is very useful in image processing where noise is transform on modified coefficients[4].
maximum and retrieval of the original image is a Basic procedure for all thresholding method is
must for proper utilization of information like in 1. Calculate DWT if the Image.
RRTT
medical image processing or in real-time video- 2. Threshold the wavelet components.
streaming where the image quality is not high and 3.Compute IDWT to obtain denoised estimate.
little noise may create large problem[9]. There are two thresholding functions frequently
IIJJEE

used i.e. Hard Threshold [10], Soft threshold. Hard-

4.2. Transform domain filtering Thresholding function keeps the input if it is larger
than the threshold; otherwise, it is set to zero. Soft-
thresholding function takes the argument and shrinks
The transform domain filtering can be further
it toward zero by the threshold. Soft-thresholding rule
divided into data adaptive and non-adaptive filters.
is chosen over hard-thresholding, for the soft-
Main type of transform domain filtering is wavelet
thresholding method yields more visually pleasant
based filtering.
images over hard thresholding. A result may still be
noisy. Large threshold alternatively, produces signal
a.Data-Adaptive Transforms
with large number of zero coefficients. This leads to
a smooth signal. So much attention must be paid to
Recently a new method called Independent select optimal threshold[4].
Component Analysis (ICA) has gained wide spread
attention.One exceptional merit of using ICA is it’s
assumption of signal to be Non-Gaussian which helps 5. Conclusion
to denoise images with Non-Gaussian as well as
Gaussian distribution[11][12]. In this paper different denoising techniques are
discussed. The study of various denoising techniques
b. Non-Adaptive Transform for digital images shows that wavelet filters’s
performance is better than spatial domain filters.
Spatial domain filters operate by smoothing over a
Wavelet transform: The term “wavelets” is
fixed window and sometimes causes over smoothing
used to refer to a set of orthonormal basis functions
and it blurred the digital image. Wavelet Transform
generated by dilation and translation of scaling

www.ijert.org 419
 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
Vol. 2 Issue 4, April - 2013

is better denoising technique than spatial filtering Research in Computer Science and Software
because of its properties like multiresolution and Engineering Research, Volume 3, Issue 2, February
multiscale nature. Main advantage of wavelet 2013 ISSN: 2277 128X
transform is that it is easy to implement. [13] A. K. Jain., “Fundamentals of digital image
processing”. Prentice Hall, Englewood Clis, NJ,
6.References c1989
[14] Gagnon, L., “Wavelet Filtering of Speckle
Noise- ome Numerical Results”, Proceedings of the
[1] M.Vijay, L.Saranya Devi, M.Shankaravadivu and
Conference Vision Interface, Trois-Reveres.
M.Santhanamari “Image Denoising Based On
Adaptive Spatial and Wavelet Thresholding [15] Goodman, J. W “Some fundamental properties
Methods”, IEEE-International Conference On of Speckle”, J. Opt. Soc. Am., 66, pp. 1145–1150.
Advances In Engineering, Science And Management, [16] Shruti Bhargava,Dr. Ajay Somkuwar,
(ICAESM -2012) March 30, 31, 2012 161 Mini Gupta, “A survey Paper on Image Denoising
[2] R.C. Gonzalez and R.E. Woods “Digital Image Techniques” International Journal of Electronics
Processing”, 2nd ed. Englewood Cliffs, NJ: Prentice- Communication and Computer Engineering, Volume
Hall; 2002. 3, Issue 3, ISSN 2249 –071X
[3] S.G.Chang, B.Yu and M.Vetterli, 2000, “Spatially
adaptive wavelet thresholding with context modeling
for image denoising,” IEEE Trans. on Image
Processing, Vol.9, No.9, pp. 1522-1531
[4] Jappreet Kaur1, Manpreet Kaur2, Poonamdeep
Kaur3, Manpreet Kaur4 “Comparative
RRTT
AnalysisofImage Denoising Techniques”
International Journal of Emerging Technology and
Advanced Engineering, ISSN 2250-2459, Volume 2,
IIJJEE

Issue 6, June 2012

[5] Motwani, M.C., Gadiya, M. C., Motwani, R.C.,
Harris, F. C Jr. “Survey of Image Denoising
Techniques”.
[6] Gleb V. Tcheslavsk, “Spatial filtering
fundamentals”, Spring 2008. ELEN 4304/5365 DIP
[7] J. Canny, "A computational approach to edge
detection," IEEE Trans.Pattern Anal. Machine Intell.,
vol. PAMI-8, pp. 679-697, Nov. 1986
[8] Bibina V C , Sanoj Viswasom “Adaptive Wavelet
Thresholding & Joint Bilateral Filtering for Image
Denoising”
[9] Chiranjib Sur “ Trilateral Filtering for Proper
Edge Smoothening and Noise Reduction”
[10] Pan, Q., “Two denoising methods by wavelet
transform”, IEEE Trans Signal Processing, vol. 47,
pp. 3401-3406.
[11] Anjali Malviya,” Fractal Based Spatial Domain
Techniques for Image De-Noising”.
[12] Rupinderpal Kaur* Rajneet Kaur, “Survey of
De-noising Methods Using Filters and Fast Wavelet
Transform”, International Journal of Advanced

www.ijert.org 420