See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/316700637

BRAIN TUMOR DETECTION USING IMAGE PROCESSING: A SURVEY

Conference Paper · November 2016

CITATION READS

1 7,005

2 authors, including:

Sudha Gupta
Somaiya Vidyavihar
24 PUBLICATIONS   104 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Machine learning View project

Smart Grid Security View project

All content following this page was uploaded by Sudha Gupta on 06 May 2017.

The user has requested enhancement of the downloaded file.

 BRAIN TUMOR DETECTION USING IMAGE PROCESSING: A
SURVEY
1
AMRUTA PRAMOD HEBLI, 2SUDHA GUPTA
1,2
Department of Electronics Engineering,
K.J.Somaiya College of Engineering,Vidyavihar, Mumbai,400077
E-mail: [email protected], [email protected]

Abstract— The Brain Tumor is affecting many people worldwide. It is not only limited with the old age people but also
detected in the early age . Brain Tumor is the abnormal growth of cell inside the brain cranium which limits the functioning
of the brain. Early detection of the brain tumor is possible with the advancement of machine learning (ML) and image
processing (IP) .In this paper stages of image processing are discussed and overview of the analogous papers are quoted by
analyzing several research papers. This paper provides gist of technologies which can be used to predict brain tumor.

Keywords — Brain Tumor, Segmentation methods, MRI.

I. INTRODUCTION

Brain tumor develops because of unusual cell growth

within the brain. Brain Tumor generally classified
into two types benign and malignant tumors[1].
Malignant Tumors are fast growing cancerous tissues.
Benign are slow growing, stagnant cancerous tumor.
Most of the tumors are life threatening , brain tumor
being one among them. Primary brain tumors
originates in the brain. In the Secondary type of brain
tumor the tumor expansion into the brain results from
other parts of the body. Imaging tumors with more
accuracy plays pivotal role in the diagnosis of tumors.
It involves high resolution techniques like MRI , CT
,PET etc. MRI is a important mean for studying the
body’s visceral structures [2]. MRI is widely used
because it gives better quality images of the brain and
cancerous tissues compared with the other medical Figure 1.Steps for image processing
imaging techniques such as X-Ray or Computed
Tomography (CT). As being a non-invasive A] Preprocessing and Enhancement of an Image
technique MRI are majorly used[13]. The basic
principle behind MRI is to generate images from MRI This is the first step of image processing it is used to
scan using strong magnetic field and radio waves of enhance the chances of detecting the suspicious
the body which helps in investigating the anatomy of region. Finer details of the image are enhanced and
the body. noise is removed from the image. Clinical MRI when
corrupted by noise reduces the accuracy of the image.
The entire paper is organized as, Existing Various filters are used to remove this noise .
methodologies in section II , comparison of K-means Anisotropic filter is used to remove background noise,
and Fuzzy clustering in section III, comparison of GA weighted median filter is used to remove salt and
and PSO in section IV, survey of different methods pepper noise. Wavelet based de-noising method
used until now are quoted in section V. makes wavelet and scaling coefficient biased. The
original image and image after enhancement is shown
II. EXISTING METHODOLOGY in figure 2.

Image Processing techniques are used to detect

tumor that has mainly following steps – Pre
Processing, segmentation, Feature Extraction and
Classification. The flowchart of the steps followed in
tumor detection and classification is shown in figure
1. Initial stage includes collection of MRI samples.
MRI has different weighted images T-1 Weighted, T-
2 Weighted and Flair -Weighted. Figure 2. Image enhancement

Proceedings of 65th IRF International Conference, 20th November, 2016, Pune, India, ISBN: 978-93-86291-38-7
76
 Brain Tumor Detection Using Image Processing: A Survey

B] Segmentation methods In this algorithm , at the start number of solutions or

Image segmentation is the method of breaking down populations are available , the solution from one
an image into small parts.segmentation is performed population is utilized to form a new population which
to make the analysis easier. There are following types is superior than the old one . As long as some
of image segmentation. condition satisfies this process is repeated.Expensive
1] Boundary approach or Thresholding computational cost is the drawback of GA.
It is the most commonly used segmentation method . d)PSO (Particle swarm optimization)
It is the gray valve remapping method where if p is It is a population based search algorithm which is
considered as an operation then as shown in equation initiated with the randomly selected populations , also
(1), called as particles. All the particles in the PSO have
their individual fitness value , which can be
p (v )= {o1 if v< t
if v≥ t
.....(1) calculated by the fitness function. They also have
velocity which directs the flying of the particle in
search space. Unlike GA ,PSO does not have a direct
where v is the gray valve and t is the threshold value . recombination operator. Drawback of PSO is , the
In the thresholding method the gray image is swarm may prematurely converge as there is rapid
converted to binary image . After thresholding the information flowing between particles. It is also
image has segmented into two values 0 and 1. problem reliant as the output rely on the parameter
2] Edge approach setting of the algorithm.Comparison between GA and
In edge-based segmentation method , the detected PSO is given in Table 2.
edges in an image are assumed to represent object C) Feature Extraction
boundaries and used to identify these objects. Edge Extracting the exact tumor is a crucial task in case of
based segmentation very rarely gives the absolute brain tumor because of the complex structure of
distinct and closed boundaries needed for a direct brain. Certain parameters are taken into account for
segmentation. Chances are more that false edge feature extraction as size , shape, composition ,
detection and many of the times it requires edge location of the image . As per the results obtained
linking to joint the partial edges into an object from the feature extraction the classification of the
boundary. tumor is done .
3] Region approach
Region based approach depends on the assumption III. COMPARISON OF K-MEANS AND FUZZY
that the bordering pixels within one region have CLUSTERING
similar values. It focuses on finding object region
instead of it's edges.It compares one pixel with its Table 1. Comparison of K-means and Fuzzy
neighbors,if the congruence criteria satisfies then the clustering
pixel can be set to belong to the cluster as one or more
of its neighbors. Different clustering algorithms are
used in this type of approach .
a)K-means algorithm
K-means algorithm is widely used clustering
technique. Which is also known as hard clustering
algorithm, it partitions a given dataset into c or k
clusters . This algorithm is simple fast and robust to
implement . It has some disadvantages as it may not
be successful to find overlapping clusters and it also
fails to cluster noisy data and non-linear datasets.
b)Fuzzy clustering
Fuzzy clustering also known as soft clustering . In this
an object is a member of a single cluster as well as a IV. COMPARISON OF GA AND PSO
member of many clusters . i.e. objects which are
located on the boundaries of the clusters are not forced Table 2. Comparison of GA and PSO
to belong to a certain cluster, rather they can be
member of many clusters. Comparison between K-
means and Fuzzy is given in Table 1.
c)Genetic Algorithms
GA and it's many versions are well received in
academe and industry because of it's advantages such
as it is ease of implementation, ability to solve higher
nonlinearities and it's intuitiveness. GA has three
main operators as recombination, mutation and
selection operator.

Proceedings of 65th IRF International Conference, 20th November, 2016, Pune, India, ISBN: 978-93-86291-38-7
77
 Brain Tumor Detection Using Image Processing: A Survey

V. SURVEY OF DIFFERENT METHODS

Table 3. Different methods used and their conclusions

Figure 3. Methodologies studied in the survey

CONCLUSION

In this work, several methodologies are examined to

denote the conventional stages of MRI image
processing also analyzed individual segmentation
approach. In conjunction with this different
methodologies proposed by the researchers are
considered to conclude that machine learning shows
an important role in brain tumor detection and
classification together with appropriate segmentation
approach. Along with this comparison between K-
means and fuzzy c, GA and PSO has also being
drafted.

REFERENCES

[1] Riddhi.S.Kapse1 , Dr. S.S. Salankar2 , Madhuri.Babar3

“Literature Survey on Detection of Brain Tumor from
MRI Images”IOSR Journal of Electronics and
Communication Engineering (IOSR-JECE) e-ISSN: 2278-
2834,p- ISSN: 2278-8735.Volume 10, Issue 1, Ver. II (Jan
- Feb. 2015), PP 80-86
[2] A.Sindhu1 , S.Meera2 “A Survey on Detecting Brain
Tumor in mri Images Using Image Processing Techniques
”International Journal of Innovative Research in Computer
and Communication EngineeringVol. 3, Issue 1, January
2015
[3] Zeynel Cebeci1 , Figen Yildiz 2 “Comparison of K-Means
and Fuzzy C-Means Algorithms on Different Cluster
Structures”Journal of Agricultural Informatics (ISSN
2061-862X) 2015 Vol. 6, No. 3:13-23
[4] Soumi Ghosh , Sanjay Kumar Dubey “Comparative
Analysis of K-Means and Fuzzy CMeans Algorithms”
((IJACSA) International Journal of Advanced Computer
Science and Applications, Vol. 4, No.4, 2013
[5] Russell C. Eberhart and Yuhui Shi “Comparison between
Genetic Algorithms and Particle Swarm Optimization ”
Indiana University Purdue University Indianapolis 723 W.
Michigan St., SL160 Indianapolis, IN 46202
[6] L. O. Hall, A. M. Bensaid, L. P. Clarkeetc, "A comparison
of neural network and fuzzy clustering techniques in
segmenting magnetic resonance images of the
brain", IEEE Trans. Neural Networks, vol. 3, no. 5, pp.
672-682, Sep. 1992.
[7] A. Kharrat, N. Benamrane, M. Ben Messaoud and M.
Abid, “Detection of Brain Tumor in Medical Images,” 3rd
International Conference on Signals, Circuits and Systems
(SCS): 1 – 6, 6-8 November 2009.
[8] J. Zhou, K. L. Chan, V. F. H. Chongand, and S. M.
Krishnan, “Extraction of Brain Tumor from MR Images
Using One-class Support Vector Machine,” Proc. IEEE

Proceedings of 65th IRF International Conference, 20th November, 2016, Pune, India, ISBN: 978-93-86291-38-7
78
 Brain Tumor Detection Using Image Processing: A Survey
Engineering in Medicine and Biology Society (EMBS 05), Support Vector Machine ” Leonardo Journal of Sciences
2005, pp. 6411-6414. ISSN 1583-0233 ,Issue 17, July-December 2010 ,p. 71-82
[9] Gopal,N.N. Karnan, M. , ―Diagnose brain tumor through [13] V. Anitha and S. Murugavalli, “Brain tumour
MRI using image processing clustering algorithms such as classification using two-tier classifier with adaptive
Fuzzy C Means along with intelligent optimization segmentation technique,” IET Computer Vision, 2016, vol.
techniques Page(s): 1 – 4, Computational Intelligence and 10(1), pp. 9-17.
Computing Research (ICCIC), 2010 IEEE International [14] Sandabad Sara; Benba Achraf; Sayd Tahri Yassine;
Conference, 28-29 Dec. 2010. Hammouch Ahmed “New method of tumor extraction
[10] M. Schmidt, I. Levner, R. Greiner, et al. Segmenting brain using a histogram study ”SAI Intelligent Systems
tumors using alignment-based features[C]. Proceedings of Conference (IntelliSys), 2015, Issue 10-11 Nov. 2015
the Fourth International Conference on Machine Learning [15] Nailah Afshan ,Shaima Qureshi,Syed Mujtiba Hussain,
and Applications, 2005, 6 pp. ”Comparative study of tumor detection
[11] Ehab F. Badran, Esraa Galal Mahmoud, and Nadder algorithms”International Conference on Medical Imaging,
Hamdy, “An Algorithm for Detecting Brain Tumors in m-Health and Emerging Communication Systems
MRI Images”. IEEE International Conference on (MedCom) 2014
Computer engineering and Systems(ICCES), pp: 368 - [16] Dilpreet Kaur, Yadwinder Kaur “Intelligent Medical
373, 2010. Image Segmentation Using FCM, GA and PSO ”
[12] Ahmed KHARRAT, Karim GASMI , Mohamed BEN International Journal of Computer Science and
MESSAOUD ,”A Hybrid Approach for Automatic Information Technologies, Vol. 5 (5) , 2014, 6089-6093.
Classification of Brain MRI Using Genetic Algorithm and



Proceedings of 65th IRF International Conference, 20th November, 2016, Pune, India, ISBN: 978-93-86291-38-7
79

View publication stats