[IJCT-V3I2P37] Authors: Amritpal Singh, Prithvipal Singh

International Journal of Computer Techniques -– Volume 3 Issue 2, Mar-Apr 2016
ISSN: 2394-2231 http://www.ijctjournal.org Page 247
Ultrasound-Image De-noising Technique’s best mix selection using
Genetic Algorithm Approach
Amritpal Singh1
, Prithvipal Singh2
1,2
(Department of Computer Science, Guru Nanak Dev University, Amritsar)
----------------------------------------************************----------------------------------
Abstract:
Medical Images are regularly of low contrast and boisterous/Noisy (absence of clarity) because of
the circumstances they are being taken. De-noising these pictures is a troublesome undertaking as they
ought to exclude any antiquities or obscuring of edges in the pictures. The Bayesian shrinkage strategy has
been chosen for thresholding in light of its sub band reliance property. The spatial space and Wavelet
based de-noising systems utilizing delicate thresholding strategy are contrasted and the proposed technique
utilizing GA (Genetic Algorithm) is used. The GA procedure is proposed in view of PSNR and results are
contrasted and existing spatial space and wavelet based de-noising separating strategies. The proposed
calculation gives improved visual clarity to diagnosing the restorative pictures. The proposed strategy in
view of GA surveys the better execution on the premise of the quantitative metric i.e PSNR (Peak Signal
to Noise-Ratio) and visual impacts. Reenactment results demonstrate that the GA based proposed
technique beats the current de-noising separating strategies.
Keywords — Speckle-noise, Medical-Imaging, Filtering-techniques, Bayesian-shrink, PSNR.
----------------------------------------************************----------------------------------
I. INTRODUCTION
In medical-imaging noise concealment alongside
the ease and cost effectiveness make ultrasound
pictures another attractive apparatus for analysis. In
the medicinal writing, speckle has been dealt with
as a diverting ancient rarity as it has a tendency to
corrupt the determination and the article
perceptibility. Ultrasound pictures are influenced by
the nearness of multiplicative noise i.e. speckle
noise that reduce the visual quality which is
acquainted due to participation of several variables
[3].
Image-Noise makes aggravation in clear visual
results. This builds the need of pre-handling the
picture by an appropriate computerized denoising
method. Speckle-Noise debases the fine points of
interest, edge location and differentiation
determination is corrupted.
Denoising of picture is a sort of picture upgrade
approach. It empowers to have a superior picture
quality.
A few condition of-workmanship denoising
separating procedures are being concentrated on by
the scientists to enhance the execution of these
channels as far as quality, decrease of run time and
numerous more elements are concerned.
II. FILTERING TECHNIQUES
The different denoising procedures are:
Lee-Filter
The Lee filter is based on the approach that is the
variance over an area is low or constant, and then
the smoothing will be performed. Otherwise, if the
variance is high (E.g. near edges), smoothing will
not performed. The main disadvantage is that it
tends to ignore speckle in the areas closest to edges
and lines [1].
RESEARCH ARTICLE OPEN ACCESS

Kuan-Filter
It does not make approximation on the noise
variance within the filter window. The main
limitation of kuan filter is that the ENL parameter is
needed for computation. The Kuan-filter is
considered to be superior to Lee filter. [1].
Wiener-Filter
Wiener is a low pass filters for handling an
intensity image which is degraded by constant
power additive noise. Wiener uses a pixel-wise
adaptive Wiener method based on statistics
estimation from a local neighborhood of each pixel.
Median-Filter
Median filter is a non-linear technique that works
best with impulse noise (salt & pepper noise) while
retaining sharp edges in the image [1].
Wavelet-Based Denoising
Donoho and Johnstone in (1994) developed a
theoretical framework for denoising signals using
Discrete Wavelet Transform (DWT). The method
consists of applying the DWT to the original data,
thresholding the detailed wavelet coefficients and
inverse transforming the set of thresholded
coefficients to obtain the denoised signal [1].
Zong et al. proposed a homomorphic wavelet
shrinkage system to partitioned the dot clamor from
the first picture [5].All the homomorphic separating
approaches by and large experience the ill effects of
two noteworthy downsides: The log change being a
nonlinear operation, prompts the one-sided
estimation of reflectivity by changing the mean of
the homogeneous regions and computationally
extremely costly because of the extra log and
exponential operations [5].
BayesShrink is a versatile information driven
edge for picture denoising by means of wavelet
delicate thresholding. Limit is driven in a Bayesian
structure, and we expect Generalized Gaussian
Distribution (GGD) for the wavelet coefficients in
every point of interest subband and attempt to
discover the edge T which minimizes the Bayesian
danger.
.
III. LITERATURE REVIEW
David and Kalyanmoy (1991), [2] considers a
number of selection schemes commonly used in
modern genetic algorithms and analysed for more
detailed analytical investigation of selection
techniques.
Hong Sern (2001), in [11] presented a new
adaptive contourlet transform-based technique for
SAR image speckles removal. The discussuion
consider the comparison of performance of Lee
filter, Forster filter, Gamma filter, wavelet-based
despeckling and contourlet transform-based
despeckling is provided for both simulated and
actual SAR images.
Lakhwinder et al. (2002), in [7] proposes an
adaptive threshold estimation method for image
denoising in the wavelet domain based on the
generalized Guassian distribution (GGD) modeling
of subband coefficients. Experimental results show
that the proposed threshold removes noise
significantly and remains within 4% of
OracleShrink and outperforms SureShrink,
BayesShrink and Wiener filtering most of the time.
Savita Gupta et al. (2003) have specified in [4]
that a novel speckle-reduction method is introduced,
based on soft thresholding of the wavelet
coefficients of a logarithmically transformed
medical ultrasound image. Experimental results
showed that the proposed method outperformed the
median filter and the homomorphic Wiener filter by
29% in terms of the coefficient of correlation and 4%
in terms of the edge preservation parameter.
S. Gupta et al. (2005) in [3] have discussed a
versatile wavelet domain despeckling technique
which visually enhances the ultrasound images for
diagnosis purpose. The different qualitative

measures are visually being compared and show the
best result.
Qi-ming, et al. (2006), [10], shows the
effectiveness of the methods which validates the
results by analysing the simulated and real signals.
Paulinas and Usinskas (2007) in [9], explains that
genetic algorithms are most powerful unbiased
optimization techniques for sampling a large
solution space.
Gupta, et al. (2007), [5] presents a versatile
wavelet domain despeckling technique to visually
enhance the medical ultrasound (US) images for
improving the clinical diagnosis. The visual
comparison of despeckled US images and the
higher values of quality metrics (coefficient of
correlation, edge preservation index, quality index,
and structural similarity index) indicate that the new
method suppresses the speckle noise well while
preserving the texture and organ surfaces.
Hashemi and Kiani (2009) in [6] demonstrated
the genetic method which was stronger than
counterpart methods in terms of contrast and detail
enhancement and producing natural looking images.
Krishna and Reddy (2010) , developed the
approach in [8] which is based on functional level
evolution whose architecture includes nonlinear
functions and uses genetic algorithm for finding the
best filter configuration.
Abassi et.al (2014) in [11] discussed blind, still
image, Genetic Programming (GP) based robust
watermark scheme for copyright protection.
IV. G.A. BASED IMAGE ENHANCEMENT
APPROACH
Genetic Algorithm (GA) accomplishes the
answers for enhancement issues regulated after the
procedure of characteristic development. Among
the evolutionary strategies, Genetic algorithms
(GAs) are the most developed gathering of
techniques that speak to the use of developmental
devices. GAs depends on the rule of "Survival of
Fittest".
GA utilizes the wonders which comprises of a
choice, hybrid and transformation administrators.
GA takes after the progressive eras to pick a
chromosome structure. An underlying populace is
produced haphazardly. Further genetic-operators,
for example, cross-over and mutation are connected
to accomplish craved advanced results [6].
Fundamentally, straightforward GA comprises of
taking after steps:
1) Initial Population produced.
2) Fitness estimation of each chromosome
taking into account some issue particular
measurements is assessed.
3) Select a couple of chromosomes on premise
of some determination system
a. Apply genetic operators i.e. cross-over
and mutation transformation
b. Selected chromosomes supplanted by
new chromosomes that are determined after
utilization of genetic
operators/administrators
4) Finally result acquired by picking the
chromosome with most noteworthy wellness
esteem.
Fundamentally, this paper dissects a picture
improvement issue which is contrasted and the
diverse denoising sifting procedures.
The proposed picture upgraded GA comprises of
chromosome structure as appeared in figure 1 & 2.
Chromosome is represented as a cluster comprising
of 6 components.
The components contain the accompanying
variables:

Fig. 1 Chromosome Structure
Fitness function
PSNR metric is used as fitness value for each
chromosome. This fitness function has been shown
in equation below:
=
where PSNR is defined as follows: Peak Signal-
to-Noise Ratio(PSNR) :
Genetic-Operators
Genetic-Operators are connected in the wake of
selecting a couple of chromosomes taking into
account Roulette Wheel Selection technique. This
technique considers the chromosomes to be a piece
of roulette wheel where every chromosome
possesses part in view of their wellness values.
Guardian chromosomes are chosen by turning the
wheel. Further crossover and mutation operators are
connected to create the best arrangement
haphazardly [10].
Crossover
Crossover operator is connected by applying
hybrid on certain arbitrary component of chose pair
of chromosomes. Any arbitrary number is chosen
from the reach 2 to 6. The chosen component
number is traded among the chosen chromosomes
as appeared underneath in figure:
Fig. 2 Crossover operation n progress
Mutation operator when applied to a selected
element of chromosome, changes its value.
Mutation point is chosen randomly between the
range 2 to 6. Mutation operator changes the values
based on element chosen:
Fig. 3 Selection of chromosome with highest fitness
.
Analysis Result
Experiments are conducted to assess the better
performance from all denoising filtering methods.

The result in table shows that GA produce better
result than Homomorphic Filtering Methods.
Wavelet based denoising algorithms uses soft
thresholding to provide smoothness and better edge
preservation but proposed GA based method
outperforms these results. The value of PSNR
should be greater than zero. Greater the value of
PSNR , better is the quality of the image. The value
of EPI and CoC should lie between 0 and 1. Value
near to 1 gives the better result. The comparison of
different filtering methods with Proposed GA is
shown in the table below
V. CONCLUSIONS
The Trial results demonstrates that the proposed
strategy accomplish preferred results over the other
homomorphic picture denoising techniques and that
too in less time (constant time multifaceted nature).
The proposed strategies chromosome contains six
parameters specifically unique picture, picture
design, clamor fluctuation, accessible denoising
techniques, wavelet deterioration sorts and
disintegration levels. Proposed GA discovers the
best mix of offered parameters to get the ideal and
close ideal results as far as time complexity is
considered. In addition when we are managing the
huge information set of boisterous pictures the
proposed procedure ensures better results with
substantially less time intricacy. This strategy can
be connected on more non specific fields of study
where numerous parameters impact a figuring for
finding ideal and close ideal results in steady time
multifaceted nature.
REFERENCES
[1] Tan H., “Denoising of Noise Speckle in Radar Image”, 2001
[2] Goldberg D. E. and Deb K., “A Comparative Analysis of Selection
Schemes Used in Genetic Algorithms”, Foundations of Genetic
Algorithms, Morgan Kaufmann, pp. 69-93, 1991.
[3] Gupta S., Chauhan R.C and Saxena S.C, “Locally adaptive wavelet
domain Bayesian Processor for denoising medical ultrasound images
using speckle modelling based on Rayleigh distribution”, IEE
Proceeding of Vision, Image and Signal Processing, vol.152, no.1, pp.
129-135,2005.
[4] Gupta S., Kaur L., Chauhan R.C, Saxena S.C, “A wavelet based
Statistical Approach for Speckle Reduction in Medical Ultrasound
Images”, IEEE Proceeding of Convergent Technologies for the Asia-
Pacific Region, Tencon, vol.2, pp. 534-537, 2003.
[5] Gupta S., Kaur L., Chauhan R.C., Saxena S.C. “a versatile technique
for visual enhancement of medical ultrasound images”, Digital Signal
Processing, vol. 17, no. 3, pp-542-560, 2007.
[6] Hashemi S., Kiani S., Noroozi N., Moghaddam M.E, “ An Image
Enhancement Method Based On Genetic Algorithm”, IEEE
Proceeding on Digital Image Processing, pp 167-171, 2009.
[7] Kaur L., Gupta S. and Chauhan R.C, “Image Denoising using Wavelet
Thresholding”, Indian Conference on Computer Vision, Graphics and
Image Processing, Ahmedabad, 2002.
[8] Krishna K.S.R., Reddy A.G., Prasad M.N.G, Rao K.C., Madhavi M.,
“Genetic Algorithm Processor for Image Noise Filtering Using
Evolvable Hardware”, International Journal of Image Processing, vol.
4, no.3, pp. 240-250, 2010.
[9] Paulinas M. and Usinskas A. , “A Survey of genetic algorithms
applications for image enhancement and segmentation” , Information
Technology and Control, vol.36, no.3, pp. 278-284, 2007.
[10] Ma Q. M., Wang X. Y., Du S. P. , “Method and application of wavelet
shrinkage denoising based on genetic algorithm”, Journal of Zhejiang
University , vol.7 no.3,pp.361-367, 2006.
[11] Abbasi A., Seng W.C., Ahmad I. S., “Multi Clock Based Image
Watermarking in Wavelet Domain using Genetic Programming”,
International Arab Journal of Information Technology, vol. 11, no. 6,
2014.

[IJCT-V3I2P37] Authors: Amritpal Singh, Prithvipal Singh

Related slideshows

More Related Content

[IJCT-V3I2P37] Authors: Amritpal Singh, Prithvipal Singh