International Journal of Engineering Trends and Technology- Volume4Issue3- 2013
Brain Tumor Identification Using K-Means
Clustering
Manali Patil
Samata Prabhu
Sonal Patil
Sunilka Patil
Mrs.Prachi Kshirsagar
I. T. Department,
Padmabhushan Vasantdada Patil Pratishthan’s College Of Engineering.
Sion (East), Mumbai-400 022, India.
Abstract
The project is entitled as “BRAIN TUMOR
IDENTIFICATION”. The idea behind choosing this topic
was to simplify the process of tumor identification. The
system will be computerized and hence time consumed
will be less. The system will also keep records of patients
who are affected by tumor and who are not. So the
doctors can schedule the further treatments for the
patients.
Keywords
Tumor, Brain, Clustering, MRI image, identifying tumor.
I. INTRODUCTION
A brain Image consists of four regions i.e. gray matter (GM),
white matter (WM), cerebrospinal fluid (CSF) and
background. These regions can be considered as four
different classes. Therefore, an input image needs to be
divided into these four classes. In order to avoid the chances
of misclassification, the outer elliptical shaped object should
be removed. By removing this object we will get rid of non
brain tissues and will be left with only soft tissues. Brain
tumor identification is used to identify tumor from particular
image. Brain tumor identification image application is
typically based on clustering concept of image pixels matrix.
Brain tumor identification is used to identify tumor affected
image based on clustering and centroid concept.
occurrence matrix approach. The level of recognition, among
three possible types of image areas: tumor, non-tumor and
back ground.
The main objective of this project is to study the design of a
computer system able to detect the presence of a tumor in the
digital images of the brain, and to accurately define its
borderlines
A. Images clustering
Convert 2-D report into 3-D images because brain is mass
body to calculate centroid so it has 3 dimensions and we need
to calculate centroid, as also we can get more accurate results
with 3 dimensions.
B. Algorithm k-means
1. Place K points into the space represented by the objects
that are being clustered.
2. These points represent initial group of centroids.
3. Assign each object to the group that has the closest
centroid.
When all objects have been assigned, recalculate the
positions of the K centroids.
Repeat Steps 2 and 3 until the centroids no longer move. This
produces a separation of the objects into groups from which
the metric to be minimized can be calculated.
The basic concept is that local textures in the images can
reveal the typical regularities of the biological structures.
Thus, the textural features have been extracted using a co-
ISSN: 2231-5381 http://www.internationaljournalssrg.org
Page 354
International Journal of Engineering Trends and Technology- Volume4Issue3- 2013
Start
II.PROPOSED SYSTEM
In our project, first the MRI report of the patient is scanned
and made into computerized form. As it becomes in
computerized form, detection of the tumor becomes simpler
as clustering is done on that MRI image and manual checking
by doctors is avoided. So the results generated are more
specific.
The major role of this application is to identify the tumor in
the brain image and reconstruct the area which the tumor is
affected and based on the threshold value the system will
identify whether the image is affected by the tumor or not.
Following are the functionality which is involved in the
tumor identification module.
No of
clusters
_
_ obj
No
move
centroid
+
end
Distance objects to centroid
Grouping based on min
distance
 Identification
 Reconstruction
 Testimony
Fig: K-means clustering
C.Testimony
A. Identification
K-Means algorithm is used to implement the Identification
of the MRI brain image. Clustering can be considered the
most important unsupervised learning problem; so, as every
other problem of this kind, it deals with finding a structure in
a collection of Unlabeled data. A loose definition of
clustering could be “the process of organizing objects into
groups whose members are similar in some way”. A cluster
is therefore a collection of objects which are “similar”
between them and are “dissimilar” to the objects belonging to
other clusters.
We are implementing the k-means algorithm with brain
image. Algorithm will cluster the brain image and
differentiate the cells into the affected cluster region and
unaffected cluster region.
B.Reconstruction
The report is generated based on the affected and the
unaffected image. The users have to select the option the
affected or un affected patients. The reports contain the
patient id and the name of the candidate.
III.STEPS OF IMPLEMENTATION
1.
2.
3.
4.
5.
Register patient details with image of his/her brain.
Select operation.
Select patient ID for identification of image.
Select clustering for image clustering.
Reconstruction of image with proper result.
A. Block Diagram
Database
The affected area will be selected and as a cluster and
constructed as an image and it is displayed in the label. Based
on the constructed area threshold values will calculated and
the tumor identification process will performed based on the
threshold values. Our system will show the option pane
message dialog contain the image affected or not.
ISSN: 2231-5381 http://www.internationaljournalssrg.org
User Login
Clustering of
3D MRI image
using threshold
value
Report
generation
Page 355
International Journal of Engineering Trends and Technology- Volume4Issue3- 2013
IV.EXPERIMENTAL RESULTS
In this paper, we identify the tumors that are located in the
brain. Here we present the system for identifying the tumors
that are presented in the brain. Brain Tumor is the tissue that
has been affected or has been damaged due to some reasons.
Brain tumors are commonly located in the posterior cranial
fossa in children and in the anterior two-thirds of the cerebral
hemispheres in adults, although they can affect any part of
the brain.
A. Tumor identification
Here, basically we are using scanned images of brain. As we
take the scanned digital image of brain, we process that
particular image n calculate the height and width and number
of pixels of that image n is converted to grayscale. Hence
after this it is converted into 3d i.e. pixels of that image is
converted to 3d n clustering concept is applied on that image.
During clustering, it was hard to decide the number of
clusters as K-means algorithm requires predefined number of
clusters. We also have to define the 1st mean value for the
clustering.
After clustering process is done, in reconstruction, we finally
get the output where we get the result whether that image
contains tumor or not.
Accuracy of finding tumor has been increased by using
automated system.
ISSN: 2231-5381 http://www.internationaljournalssrg.org
Page 356
International Journal of Engineering Trends and Technology- Volume4Issue3- 2013
The design of the database is flexible ensuring that the
system can be implemented. It is implemented and gone
through all validation.
VI.REFERENCES
[1]. Gerig, G., Martin, J., Kikinis, R., Kubler, O., Shenton,
M., Jolesz, F.: Automating Identification of dual-echo MR
head data. In: IPMI.
Volume 511. (1991) 175{185
[2]. Kjaer, L., Ring, P., Thomson, C., Henriksen, O.: Texture
analysis in quantitative MR imaging: Tissue characterization
of normal brain
and intracranial tumors at 1.5 T. Acta Radiologic (1995)
[3]. War_eld, S., Dengler, J., Zaers, J., Guttman, C., Wells,
W., Ettinger, G., Hiller, J., Kikinis, R.: Automatic
identi_cation of gray matter
structures from MRI to improve the Identification of white
matter lesions. Journal of Image Guided Surgery 1 (1995)
326{338
[4]. Vinitski, S., Gonzales, C., Mohamed, F., Iwanaga, T.,
Knobler, R., Khalili, K., Mack, J.: Improved intracranial
lesion characterization
by tissue Identification based on a 3D feature map. Mag Re
Med (1997) 457{469
[5]. Zhu, Y., Yan, H.: Computerized tumor boundary
detection using a hop_eld neural network. IEEE-TMI (1997)
55{67
[6]. Dickson, S., Thomas, B.: Using neural networks to
automatically detect brain tumours in MR images. Int J
Neural Syst (1997) 91{99
V.CONCLUSION
[7]. Just, M., Thelen, M.: Tissue characterization with T1,
T2, and proton density values: Results in 160 patients with
brain tumors.
The “BRAIN TUMOUR IDENTIFICATION” has been
developed to satisfy all proposed requirements.
[8]. O’reilly, Java Swings, Tata McGraw Hill, Fifth Edition,
2002
The system is highly scalable and user friendly. Almost all
the system objectives have been met. The system has been
tested under all criteria. The system minimizes the problem
arising in the existing manual system and it eliminates the
human errors to zero level.
[9]. Java Handbook, Patrick Naughton & Michael Morrison,
McGraw Hill, 0-078-82199-1,1996
[10]. The Java AWT Reference, O'Reilly & Associates, Inc. ,
John Zukowski,1-565-92240-9,1997
ISSN: 2231-5381 http://www.internationaljournalssrg.org
Page 357