MIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLS

International Journal of Innovative Research in Advanced Engineering (IJIRAE)
Volume 1 Issue 1 (April 2014)
__________________________________________________________________________________________________
ISSN: 2278-2311 IJIRAE | http://ijirae.com
© 2014, IJIRAE All Rights Reserved Page - 164
MIP AND UNSUPERVISED CLUSTERING FOR THE
DETECTION OF BRAIN TUMOUR CELLS
Mrs.Sujatha.K* Mrs. Udaya Rani. V Dr.Vinayakamurthy.M
PG Scholar/CSE, Associate Professor /CSE Professor /MCA
Reva ITM,Bangalore. Reva ITM,Bangalore. Reva ITM, Bangalore.
sujathasjcit@gmail.com udayamurthy@yahoo.com dr.m.vinayakamurthy@gmail.com
ABSTRACT:-Image processing is widely used in biomedical applications. Image processing can be used to analyze
different MRI brain images in order to get the abnormality in the image .The objective is to extract meaningful
information from the imaged signals. Image segmentation is a process of partitioning an image in to different parts.
The division in to parts is often based on the characteristics of the pixels in the image. In our paper the segmentation
of the tumour tissues is carried out using k-means and fuzzy c-means clustering.Tumour can be found and faster
detection is achieved with only few seconds for execution. The input image of the brain is taken from the available
database and the presence of tumourin input image can be detected.
Keywords: Morphological Image Processing (MIP), Image Segmentation, K-Means, Fuzzy C-Means
I.INTRODUCTION
Brain tumour is an abnormal growth of the cells inside the brain. Tumour can be cancerous or noncancerous. It is
generally caused by abnormal and uncontrolled cell division which is normally in the brain itself or in the cranial nerves,
or in the brain envelopes, skull, pituitary glands or spread from cancer primarily located on other organs. [1].Image
Processing can be used to analyse different medical and MRI images to get the abnormality in the image. Medical Image
segmentation deals with segmentation [9] of tumour in MRI images for improved quality in medical diagnosis. Magnetic
resonance imaging (MRI) is an advanced medical imaging used to produce high resolution images of the parts contained
in the human body. MRI is used when treating brain tumour. These high resolution images are used to examine human
brain development and discover abnormalities. Morphological filtering of a binary image is conducted by considering
operations like opening and closing. K-means clustering is suitable for biomedical image segmentation that solves the
well-known clustering problem. It is one of the simplest unsupervised learning algorithms. Segmentation is carried out
using K-means clustering algorithm for better performance. This paper proposes automatic method to find the tumour
cells using morphological technique. It is a tool to extract the region of interest among the image. [4] Image compression
is a technique that removes the pixel redundancy and compresses the image without any loss of information. We need
compression we use JPEG 2000 standard for compressing with no error. Fig1 shows an example of segmentation, where
the original image is segmented based on individual surfaces, objects texture, boundary. Image segmentation has wide
applications like image database, image compression, recognition of objects, boundary calculations etc. The aim of
segmentation is to change the representation of an image which is simpler, easier and meaningful to analyze
Fig.1 Example of segmentation
II.METHODOLOGY
Clustering is used in image segmentation where data is organized in to groups (clusters) such that the data objects that are
similar to each other are put in the same cluster. Clustering is a form of unsupervised [5] learning in which no class labels
are provided but data records need to be grouped based on how similar they are to other records.
A.Image Enhancement
Image enhancement is a process of adjusting digital images so that the results are more suitable for display or further
image analysis .it improves the qualities of an image.

__________________________________________________________________________________________________
Fig.2 Image enhancement
Fig2 shows image enhancement where the input image is enhanced in order to get a better image.
B.Morphological operations
Morphological image processing is a collection of nonlinear operations related to the shape or morphology of features in
an image. Morphological filtering of a binary image is conducted by considering operations like opening and closing.
The morphological operations are applied to gray scale images to segment the tumour [2, 3]. Morphological process is
conducted to extract the required region. The use of this operation is to show only the part of the image which has tumour
which is identified as white color that is specified using the operation strel command, which acts as a morphological
structuring element.
C. Proposed method
In the detailed level design the details and flowchart of each of the module has been described .This involves module
specification such as
Brain Image from MRI scan
Image enhancement
Morphological operations
Fuzzy C-Means clustering
K-Means clustering
Performance analysis
Fig3 explains the flowchart of the proposed method the input to the system would be MRI images of the brain. Image
enhancement is carried out to produce the clear visual of the image after the enhancement of the image morphological
process is carried out to extract the required region. Next by implementing the two unsupervised algorithms K-means and
Fuzzy C-Means with clusters the exact result is produced. In the proposed approach we combine both segmentation and
clustering. Due to the unsupervised nature the proposed method is efficient and less error sensitive. At the end the output
is image where tumours are segmented and can be differentiated from other parts of the brain.
Fig.3 Proposed Method
D. Sequence Diagram
Fig4 shows the sequence diagram which defines between user input brain images and by sequences it works and finally
gives the output result. The input to the system is the original images from MRI scan of a brain, image enhancement is
carried out to get clear representation of an image and then the morphological operations are carried out through
processes like erosion, dilation, opening and closing. Next by implementing K-means and Fuzzy C-means with clusters
the tumors are segmented and the final result is obtained.

__________________________________________________________________________________________________
Fig.4 Sequence Diagram
III .PROPOSED ALGORITHM
Segmentation
Segmentation is a process of partitioning an image in to different parts. The division in to parts is often based on the
characteristics of the pixels in the image. In our paper the segmentation of the tumour tissues is carried out using k-means
and fuzzy c-means clustering.
K-Means based segmentation
The k-means clustering [6, 7] was introduced by MacQueen.it is one the simplest unsupervised learning algorithm that
solves the well-known clustering problem.
The procedure follows a simple and easy way to classify a given set through a certain number of clusters i.e.; K (fixed)
The k-means based segmentation depends on centroid, the center of mass of geometric objects. The centroid is based on
minimum distance.
The k-means is used for image retrieval algorithms, is traditional and just needs to do distance calculation.
1. The idea is to define k centroids one for each cluster. This centroid should be placed in a cunning way because
of different locations causes different results. so the better choice is to place them as much as possible far away
from each other.
2. Take each point belonging to a given data set and associate it to the nearest centroid when no point is pending
the first step is completed and early group is done. At this point we need to recalculate k new centroids as
centers of the clusters resulting from the previous step.
3. We have k new centroids and new binding has to be done between the same data set points and the nearest new
centroids. A loop has been generated as a result of this loop we may notice that the k centroids change their
location step by step until no more changes are done .In other words centroids do not move any more.
Fuzzy C-Means Based Segmentation
A modification by Bezdek[8]of an original crisp clustering method. Bezdek introduced the idea of fuzzification
parameter i.e. (M) in the range [1, N] which determines the degree of fuzziness in the clusters when M=1 the effect is a
crisp clustering of points. When M>1 the degree of fuzziness among points in the decision space increases.
Here the number of clusters and the fuzzification parameter is fixed.
1. Randomly initialize the cluster center.
2. Creating distance matrix from a point xi to each of the cluster centers to with taking the Euclidean distance
between the points and the cluster centers.
d=√∑ (xi- ci)
2
3. Creating membership matrix takes the fractional distance from the points to the cluster centers and makes this a
fuzzy[10] measurement by raising the fraction to the inverse fuzzification parameter this is divided by the same
of all fractional distances there by ensuring that the sum of all membership is 1.
4. Creating membership matrix fuzzy c-means imposes a direct constraint on the fuzzy membership functions
associated with each point as follows the total membership for a point in sample or decision space must add to 1.
5. Generating new centroid for each cluster.
6. Generating new centroid for each cluster with iterations all this step optimize cluster centers will generate.
IV. RESULTS AND DISCUSSION
Input to the system would be MRI images of the brain which are present in database. Output is where tumours are
segmented and can be differentiated from other parts of the brain. After the execution of fuzzy c-means and k-means
clustering algorithm the output of the GUI is given below

__________________________________________________________________________________________________
FIG 5 IMAGES OF THE BRAIN FROM MRI SCAN
Fig5 shows the select test images from the database where the images of the brain from MRI scan are stored.
FIG 6MAIN GUI WINDOW
Fig 6 shows the representation of a main graphical user interface window.
FIG 7RESULT WINDOW AFTER EXECUTION OF FUZZY C-MEANS CLUSTERING
Fig 7 shows the output after execution of fuzzy c-means clustering where the tumors are segmented.
The proposed method has been successfully implemented and tested with wide range of brain images. The tumours are
segmented using k-means and fuzzy c-means clustering, but the segmentation is faster using k-means when compared to
tumour segmentation using fuzzy c-means clustering.

__________________________________________________________________________________________________
FIG 8 RESULT WINDOW AFTER EXECUTION OF K-MEANS CLUSTERING
Fig 8 shows the output after execution of k-means clustering where the tumors are segmented
FIG 9 COMPARISON OF K-MEANS AND FUZZY C-MEANS CLUSTERING
V.CONCLUSION
Segmentation algorithms used were k-means and fuzzy c-means which made segmentation work easier. Samples of
human brains were taken which were scanned by using MRI process and then applying k-means and fuzzy c-means
algorithms the results were processed with less execution time. Morphological techniques with segmentation methods
were been used and the tumours were segmented. Unsupervised segmentation is better than supervised segmentation
methods, because supervision is not required.
REFERENCES
[1] PratibhaSharma, manojdiwakar,sangamchoudhary “Application of egde detection for brain tumour detection” volume 58
No2(2011).
[2] Digital image processing-Rafael C.Gonzalez,RichardE.woods-ADDISON-WESLEY,an imprint of pearson education ,1st edition.
[3] J.Goldberger,S.Gordon,H.Greenspan, ”Unsupervised Image-set clustering using an information theoretic framework”,IEEE
Transactions on image processing,vol 15,No 2,February 2006.
[4] Morphological image processing approach on the detection of tumour and cancer cells ICDCS,2012 International conference on
march 15-16.
[5] S.R.Kannan,jan 2005”segmentation of MRI using new Unsupervised Fuzzy C means algorithm “ICGST-GVIP jornal,Volume
5,Issue2.
[6] Kharrat, A.Benamrane, N.BenMessaoud, M.Abid, Comput. &Embedded Syst Lab. (CES),Net.Eng.Sch. ofSfax, Sfax, Tunisia,
Nov,2009.
[7] Fahim.A.M ,Salem A.M, Torkey F.A, Ramadan M.A “An Efficient enhanced K-Means Clustering Algorithm” Journal of
Zhejiang University Science A 2006
[8] Juraj Horvath,2006 ”Image segmentation using Fuzzy C-Means” SAMI 206
[9] Y.Zhang,”A Survey on evaluation methods for image segmentation “,Pattern Recognition,vol.29,pp.1335-1346,1996.
[10] S.R.Kannan,jan 2005”segmentation of MRI using new Unsupervised Fuzzy C means algorithm “ICGST-GVIP jornal,Volume
5,Issue2.

MIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLS

Related slideshows

More Related Content

MIP AND UNSUPERVISED CLUSTERING FOR THE DETECTION OF BRAIN TUMOUR CELLS