IRJET- Image Processing for Brain Tumor Segmentation and Classification

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 12 | Dec 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.34 | ISO 9001:2008 Certified Journal | Page 388
Image Processing for Brain Tumor Segmentation and Classification
Miss. Roshani S. Thombare1, Mr. Girish D. Bonde2
1M.Tech Student 2Assistant Professor, Department of Electronics and Telecommunication Engineering, J.T.M.
College of Engineering, Faizpur, India
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - In this paper, we have fostered a new transpire for
automatic dissection of brain tumors in MR images. The brain
tumor segmentation is performed using watershedtransform.
The propositioned method for brain tumor classification
entails of four stages explicitly pre-processing, DWT feature
extraction, principalcomponentanalysisforfeaturereduction,
feature extraction and classification. In pre-processing
adaptive histogram equalization is used for noise reduction
and to render the image apposite for extracting the features.
In the second stage, DWT features are extracted from the
image. In the third stage, Principal Component Analysis(PCA)
is depleted to demote the dimensionality of the feature space
which results in a more efficient and accurateclassification. In
the feature extraction stage different texture and statistical
features such as contrast, correlation, energy, homogeneity,
kurtosis, energy, and entropy are extracted. Finally, in the
classification stage, we have evaluated the performance of
SVM, KNN and neural network classifiers.
Key Words: Tumor Segmentation, DWT, PCA, SVM
Classification, KNN, Neural Networks
1. INTRODUCTION
The Brain is one of the most vital and complex structure in a
human body. Brain can be affected from problem which
cause change in its normal structure and its normal
behavior. This problem is known as brain tumor. Different
image techniques like Magnetic Resonance Imaging (MRI)
are used by inner structure of the brain and diagnose the
tumor. A tumor is an any mass cause abnormal and
uncontrolled growth of cell. There are three type of Brain
tumors have been classified namely Giloma Tumors,
Malignant Tumors, Pituitary Brain Tumor.
Benign tumor is having their boundaries or the edges. They
do not spread over the other parts of the body. Malignant
tumor is considered to be the most serious one and they
develop rapidly. They affect the various necessary organs
which even lead to the death.
Brain Tumor is includes Computed Tomography (CT) scan,
Magnetic Resonance Imaging(MRI)scan.Classificationisthe
process in which the brain image is classified as Benign or
Malignant. In this process, different features are extracted.
The classification process is very important. The
segmentation process is to perform to the system to detect
accurate tumor in image.
Segmentation is used to dividing the image to its parts
sharing identical properties like color, texture, contrast and
boundaries. Analysis of segmentation of a largesetofimages
and comparisons of these segmentations between relevant
subgroups of images. Adaptive histogram thresholding is
implemented for segmentation.
The proposed system suggests using three stages namely
Feature Extraction, Feature Reduction and Classification.
This approach extracts features using Discrete Wavelet
Transform and Feature Reduction is done using Principal
Component Analysis (PCA). Different texture and statistical
features are extracted. The result is used in Support Vector
Machine (SVM) for tumor classification as Benign or
Malignant or Pituitary. Wavelet transform is an executive
tool for feature extraction from MR brain images. The
following step as follows:
Step 1: Input of MRI Image.
Step 2: Generate Preprocessing image.
Step 3: Segmentation is locate the object and pixel image.
Step 4: Feature extraction through DWT
Step 5: Feature reduction through PCA.
Step 6: Texture and statistical features extraction
Step 6: Finally SVM classifier is used to classification from MR
Image. It is shown in Figure 1.
This paper is structured as follows: Section 2 relates with
related work done in the brain tumor segmentation and
classification. Section 3 gives the step by step method
followed for classification and segmentation. Section 4
analyses and discusses the results obtained and Section 5
concludes the work done.
Figure 1 Block Diagram of Proposed System
2. RELATED WORK
Saleck et al [4] introduced a new approach using FCM
algorithm, in order to extract the mass from region-of
Input Image
Preprocessing
Image Segmentation
Feature Extraction
Classification

interested (ROI). The proposed method aims at avoiding
problematic of the estimation of the cluster number in FCM
by selecting as input data, the set of pixels which are able to
provide us the information required to perform the mass
segmentation by fixing two clusters only. The Gray Level
Occurrence Matrix (GLCM) is used to extract the texture
features for getting the optimal threshold, which separate
between selected set and the other sets of the pixels that
influences on the mass boundaryaccuracy.Theperformance
of the proposed method is evaluated by specificity,
sensitivity and accuracy.
Bhima and Jagan [6] demonstrated the superioraccuracyfor
brain tumor detection in compared to the presented
methodologies. Also the major identified bottleneck of the
recent research outcomes are limited to detection of brain
tumor and the overall analyses of internal structure of the
brain is mostly ignored being one of the most important
factor for disorder detection.
Vrji and Jayakumari [7] improved brain tumor
approximation after a manual segmentation procedure and
2D & 3D visualization for surgical planning and assessing
tumor. The tumor identification, the investigations hasbeen
made for the potential use of MRI data for improving brain
tumor shape approximation. In Preprocessing and
Enhancement stage, medical image is converted into
standard formatted image. Segmentation subdivides an
image into its constituent regions or objects.
Rashid et al [8] investigated the chosen brain MRIimageand
a method is targeted for more clear view of the location
attacked by tumor. An MRI abnormal brain images as input
in the introduced method, Anisotropic filtering for noise
removal, SVM classifier for segmentationandmorphological
operations for separating the affected area from normal one
are the key stages if the presented method. Attaining clear
MRI images of the brain are the base of this method. The
classification of the intensities of the pixels on the filtered
image identifies the tumor.
Sudharani et al [9] the present paper proposed the
c1assification and identification scores of brain tumor by
using k-NN algorithm which is based on training of k. In this
work Manhattan metric has applied and calculated the
distance of the c1assifier. The algorithm has been
implemented using the Lab View.
Vidyarthi, A., &Mittal [10] proposed a hybrid model which
identifies the region of interest using fused results of
threshold segmentation and morphological operations.
Initially, an abnormal brain MR imageisprocessedwithOtsu
threshold basedsegmentationandmorphological operations
like erosion. Further, both the segmented resultant images
are fused with the original MR image to preserve the
background and correctly identification of thetumor region.
Li et al [11] proposed framework employs local binary
patterns (LBPs) to extract local image features, such as
edges, corners, and spots. Two levels of fusion (i.e., feature-
level fusion and decision-level fusion) are applied to the
extracted LBP features along with global Gabor featuresand
original spectral features, where feature-level fusion
involves concatenation of multiple features before the
pattern classification process while decision-level fusion
performs on probability outputs of each individual
classification pipeline and soft-decision fusion rule is
adopted to merge results from the classifier ensemble.
Moreover, the efficientextremelearningmachine witha very
simple structure is employed as the classifier.
Dhanaseely et al [12] presented and investigated two
different architecturesinthiswork.Thecascadearchitecture
(CASNN) and feed forward neural architecture (FFNN) are
investigated. The feature extraction is performed using
principal component analysis (PCA) as it reduces the
computational burden.
For a given database the features are extracted using PCA.
The Olivetti Research Lab (ORL) database is used. The
extracted features are divided into training set and testing
set. The training data set is used to train both the neural
network architectures. Both are tested extensively using
testing data.
Liu and Liu [13] proposed an algorithm of HV microscopic
image feature extraction and recognitionusinggraylevel co-
occurrence matrix (GLCM) in order to effectively extract the
feature information of human viruses (HV) microscopic
images. Firstly, 20 pieces of microscopic images of human
virus are obtained by using GLCM, and then the four texture
feature parameters, entropy, energy inertia moment and
correlation are extracted utilizing the GLCM, and then HV
image recognition is carried out.
Parveen and Singh [14] proposed a new hybrid technique
based on the support vector machine (SVM) and fuzzy c-
means for brain tumor classification.
3. BRAIN TUMOR SEGMENTATIONANDCLASSIFICATION
In this study we have used brain tumor dataset containing
3064 T1-weighted contrast-enhanced images from 233
patients with three kinds of brain tumor: meningioma (708
slices), glioma (1426 slices), andpituitarytumor(930slices)
Figure 2 shows brain tumor images from the dataset.
(a) (b)
Figure 2 MRI scan image with brain tumor
A. Brain Tumor Segmentation
Segmentation using the watershed technique works well if

the foreground objects and background regions are
identified or marked. It is a simple, perceptive methodandit
is fast. Watershed Division Method extracts seeds, which
indicate the nearness of items or foundation at particular
areas. There Marker areas are thensetto bethelocal minima
inside the topological surface and the watershed calculation
is connected. The advantage of watershed division is that it
creates an extraordinary answer for a specific picture input.
Figure 3 shows the brain tumor segmentation using
watershed transform.
Figure 3 Brain Tumor Segmentation
B.Brain Tumor Classification
There are four steps of Brain Tumor Classification:
a) Segmentation,
b) Feature Extraction by Using DWT,
c) Feature Reduction by Using PCA,
d) Statistical and Texture Features,
e) Tumor detection and classification.
a) Threshold segmentation
Image segmentation is typically used to locate objects and
boundaries such as lines, curves, etc., in images. Each of the
pixels in a region is have some similar characteristic or
computed property, such as color, intensity, or texture.
Thresholding technique segmentstheMR images bya binary
partitioning of the image intensity. The segmentation is
based on thresholds. Usually MRI images have non-isotropic
vowel sizes in order to get isotropic image, compute the
threshold of different tissues gray levels and the image
histogram as a probability density function of the Image.
b) The Discrete Wavelet Transform (DWT) became a very
versatile signal processing tool proposed the
multi‐resolution representation of signals based on wavelet
decomposition. Feature Extraction is used to extract the
wavelet coefficient from MR images. DWT is a technique
used to extract features of each imagefrombrainMRI, which
extracts maximum highlighting pixels present in images to
progress results. The main advantageof waveletsisthatthey
provide localized frequency information of classification. In
this various statistical features are calculated. There are 1)
Mean, 2) Standard Deviation, 3) Entropy.
Figure 4 DWT schematically
It is shown in Figure4. The input image is process by F1 [n]
and F2 [n] filters which is the row representation of the
original image. A result of this transform there are 4 sub
band (LL, LH, HH, HL) images at each scale.
c) Feature reduction by using principal component analysis
(pca)
The Principal Component Analysis well- accredited toolsfor
transmuting the presented input features into a new lower
dimensional feature space. In proposed method is extract
features from brain MR images using PCA. It is commonly
used forms of dimensional reduction. The main purpose of
using PCA approach is to reduce the dimensionality of the
wavelet coefficients.
Algorithm
Step1: Input Image I [x, y].
Step2: Arrange in two column vector.
Step3: Calculate Empirical means values are subtracted.
Step4: Find the covariance Matrix.
Step5: Compute Eigen vector and Eigen valuesofresultvector.
d) Statistical and Texture Features
After the segmentation is performed on tumor region, the
segmented modules are used for feature extraction. Feature
extraction is one of the most important steps in this system.
A feature is a significant piece of information extracted from
an image which provides moredetailedunderstandingof the
image. A feature is defined as a function of one or more
measurements, the values of some quantifiable property of
an object, computed so that it quantifies some significant
characteristics of the object.
i. GLCM Features
A gray level co-occurrence matrix is a second order
statistical measure. GLCM is the gray-level co-occurrence
matrix (GLCM), also known as the gray level spatial
dependence matrix. The Gray-Level Co-occurrence Matrix

(GLCM) is based on the extraction of a gray-scale image. The
GLCM functions characterize the texture of an image by
calculating how often pairs of pixel with specific values and
in a specified spatial relationship occur in an image, creating
a GLCM, and then extracting statistical measures from this
matrix. Statistical parameters calculated from GLCM values
are as follows:
The matlab function glcm = graycomatrix(I) creates a gray-
level co-occurrence matrix (GLCM) from image I.
graycomatrix creates the GLCM by calculating how often a
pixel with gray-level (grayscale intensity) value i occurs
horizontally adjacent to a pixel with the value j. (You can
specify other pixel spatial relationships using the 'Offsets'
parameter -- see Parameters.) Each element (i, j) in glcm
specifies the number of times that the pixel with value i
occurred horizontally adjacent to a pixel with value j.
The matlab function graycomatrix calculates the GLCMfrom
a scaled version of the image. By default, if I is a binary
image, graycomatrix scales the image to two gray-levels. If I
is an intensity image, graycomatrix scales the image to eight
gray-levels.
stats = graycoprops(glcm, properties) calculates the
statistics specified in properties from the gray-level co-
occurence matrix glcm. glcm is an m-by-n-by-parrayofvalid
gray-level co-occurrence matrices. If glcm is an array of
GLCMs, stats is an array of statistics for each glcm.
graycoprops normalizes the gray-level co-occurrencematrix
(GLCM) so that the sum of its elements is equal to 1.
Each element (r,c) in the normalized GLCM is the joint
probability occurrence of pixel pairs with a defined spatial
relationship having gray level values r and c in the image.
graycoprops uses the normalized GLCM to calculate
properties.
Table 1 GLCM properties
Property Description Formula
'Contrast' Returns a
measure of the
intensity
contrast
between a pixel
and its
neighbor over
the whole
image.
Range = [0
(size(GLCM,1)-
1)^2]
Contrast is 0
for a constant
image.
Property Description Formula
'Correlation' Returns a
measure of
how correlated
a pixel is to its
neighbor over
the whole
image.
Range = [-1 1]
Correlation is 1
or -1 for a
perfectly
positively or
negatively
correlated
image.
Correlation is
NaN for a
constant image.
'Energy' Returns the
sum of squared
elements in the
GLCM.
Range = [0 1]
Energy is 1 for
a constant
image.
'Homogeneity' Returns a value
that measures
the closeness of
the distribution
of elements in
the GLCM to
the GLCM
diagonal.
Range = [0 1]
Homogeneity is
1 for a diagonal
GLCM.
Along with GLCM following features are extracted
ii. Mean
M = mean (A) returns the mean values of the elements along
different dimensions of an array. If A is a vector, mean(A)
returns the mean value of A. If A is a matrix, mean(A) treats
the columns of A as vectors, returning a row vector of mean
values.
iii. Standard Deviation
s = std (X), where X is a vector, returns the standard
deviation using (1) above. The result s is the square root of
an unbiased estimator of the varianceofthepopulationfrom
which X is drawn, as long as X consists of independent,

identically distributed samples. If X is a matrix, std (X)
returns a row vector containingthestandarddeviationofthe
elements of each column of X. If X is a multidimensional
array, std(X) is the standard deviation of the elements along
the first non-singleton dimension of X.
iv. Entropy
E = entropy(I) returns E, a scalar value representing the
entropy of grayscale image I. Entropy is a statistical measure
of randomness that can be usedtocharacterizethetextureof
the input image. Entropy is defined as
-sum (p.*log2 (p))
Where p contains the histogram counts returned from
imhist. By default, entropy uses two bins for logical arrays
and 256 bins for uint8, uint16, or double arrays.
I can be a multidimensional image. If I have more than two
dimensions, the entropy function treats it as a
multidimensional grayscale image and not as an RGB image.
v. RMS
The RMS block computes the true root mean square (RMS)
value of the input signal. The true RMS value of the input
signal is calculated over a running average window of one
cycle of the specified fundamental frequency
vi. Varience
V = var(X) returns the variance of X for vectors.Formatrices,
var(X)is a row vector containing thevarianceof eachcolumn
of X. For N-dimensional arrays, var operates along the first
nonsingleton dimension of X. The result V is an unbiased
estimator of the variance of the population from which X is
drawn, as long as X consists of independent, identically
distributed samples.var normalizes V by N – 1ifN >1,where
N is the sample size. This is an unbiased estimator of the
variance of the population from which X is drawn, as long as
X consists of independent, identically distributed samples.
For N = 1, V is normalized by 1.
vii. Kurtosis
k = kurtosis(X) returns the sample kurtosis of X. For vectors,
kurtosis(x) is the kurtosis of the elements inthevector x. For
matrices kurtosis(X) returns the sample kurtosis for each
column of X. For N-dimensional arrays, kurtosis operates
along the first nonsingleton dimension of X.
viii. Skewness
y = skewness(X) returns the sample skewness of X. For
vectors, skewness(x) is the skewness of the elements of x.
For matrices, skewness(X) is a row vector containing the
sample skewness of each column. For N-dimensional arrays,
skewness operates along the firstnonsingletondimension of
X.
e) Classification using support vector machine (SVM)
SVM is a supervised machine learning algorithm. It can be
used for classification and regression. It is mostly used in
classification problems. It is state of the art pattern
recognition technique grown up from statistical learning
theory.
The basic idea of applying SVM for solving classification
problems: a) through a non-linear mapping function to
transform the input spaceto higherdimensionfeaturespace.
b) The separating hyper plane with maximum distancefrom
the closest points of the training set will be constructed.
Using a SVM classifier to accurately classify featuresof every
image in dataset.
The last step for diagnosis and which is specifically used for
classification of tumors. Therearetwoimportantcommands
svm train and svm classify used for this process. In Kernel
space the training data to map by using Kernel function svm
train. The kernel function can be following the function:
linear, quadratic, polynomial. SVM Strut is used to classify
each row of the data in sample, a matrix ofdata andusingthe
information, created using the svm train function.
4. RESULT ANALYSIS
A. Brain Tumor Segmentation
The experiments areconductedonthe braintumordetection
system where the inputs are MRI images ofbrain.MRIimage
is successfully processed at each step in brain tumor
detection system and the desired result is obtained. MRI
image of brain is given to watershed transform. For the
purpose of image segmentation marker- controlled
watershed segmentation is used. Resultant output from the
segmentationmethodaregeneratedandevaluated.Obtained
results are shown in Figure 5.
Figure 5 Brain Tumor Segmentation

B. Brain tumor Classification
The MATLAB based GUI designed for brain tumor
classification is depicted in figure 6.
Figure 6 Brain Tumor Classification
In this study we have used brain tumor dataset containing
3064 T1-weighted contrast-enhanced images from 233
patients with three kinds of brain tumor: meningioma (708
slices), glioma (1426 slices) and pituitary tumor (930slices)
The performance is evaluated using SVM, KNN and Neural
networks classifiers.
C. Accuracy arithmetic
MATLAB is a tool for the analysis and check the efficiency of
the algorithm and proposed model. The algorithm’s
performance can be evaluated in terms of accuracy,
sensitivity, and specificity. Theconfusionmatrixdefining the
terms TP, TN, FP, and FN from the predicted class and actual
class result for the calculation of accuracy as shown in Table
2
Table 2 Confusion matrix defining the term TN, FP, FN and
TN
Where,
TP is the number of true positives, which is used to indicate
the total number of abnormal cases correctly classified,
TN is the number of true negatives, which is used to indicate
normal cases correctly classified;
FP is the number of false positive, and it is used to indicate
wrongly detected or classified abnormal cases; when they
are actually normal cases and
FN is the number of false negatives; it is used to indicate
wrongly classified or detected normal cases; when they are
actually abnormal cases.
All of these outcome parameters are calculated using the
total number of samples examined for the detection of the
tumor. The quality rate parameter accuracy is definedasthe
ratio of all the correct predictions dividedbytotal numberof
cases examined. Formula to calculate accuracy is given
below.
The test performance of the classifier determined by the
computation of the statistical parameterssuchasaccuracyin
comparison with different classifier techniques is shown in
Table 3. Furthermore, higher values of accuracy indicate
better performance.
Table 3 illustrates the comparison of various classifier
techniques with accuracytheproposedsystemgives76.39%
accuracy for Neural Network, 65.61% for k-Nearest
Neighbor and 60.85% for Support Vector Machine. The
detailed analysis of performancemeasuresisshowninTable
3 and through the performance measure, it is depicted that
the performance of the proposed methodology has
significantly improved the tumor identification compared
with the NNs, SVM and 𝐾-NN basedclassificationtechniques.
Table 3 Comparison of accuracies in different classifier
Number of Test Images(919)
Evaluation
Parameter
NNs SVM K-NN
True negative 538 363 479
False positive 128 137 152
True positive 164 187 124
False negative 89 232 164
Accuracy (%) 76.39 60.85 65.61

5. CONCLUSIONS
In this paper, the concepts related to image segmentation
and Classification of brain tumor detection is discussed.
The watershed based thresholding approach is used for
image segmentation which helps to recognize the portion of
tumor in MRI image. Then using the approach of Discrete
Wavelet Transform for Feature Extraction and Reduction of
Principle Component Analysis, The Features Extracted are
Mean, Standard deviation, Kurtosis, Skewness, Entropy,
Contrast, Variance, Smoothness,CorrelationandEnergy. The
accuracy of the brain tumor segmentation is also measured.
Using this algorithm the brain tumors are accurately
segmented and classification from an MR brain image.Then,
Support Vector Machine is used to classify the tumors into
benign, pituitary and malignant.
REFERENCES
[1] Borole, V. Y., Nimbhore, S. S., & Kawthekar, D. S. S.
(2015). Image Processing Techniques for Brain Tumor
Detection: A Review. International Journal of Emerging
Trends & Technology in Computer Science (IJETTCS),
4(5), 2.
[2] Bahadure, N. B., Ray, A. K., &Thethi, H. P. (2017). Image
analysis for MRI based brain tumor detection and
feature extraction using biologically inspired BWT and
SVM. International journal of biomedical imaging, 2017
[3] Coatrieux, G., Huang, H., Shu, H., Luo, L., & Roux, C.
(2013). A watermarking-based medical image integrity
control system and an image moment signature for
tampering characterization. IEEE journal of biomedical
and health informatics, 17(6), 1057-1067.
[4] Saleck, M. M., ElMoutaouakkil, A., &Mouçouf, M. (2017,
May). Tumor Detection in Mammography Images Using
Fuzzy C-means and GLCM Texture Features. In
Computer Graphics, Imaging and Visualization, 2017
14th International Conference on (pp. 122-125). IEEE.
[5] Zulpe, N., &Pawar, V. (2012). GLCM textural features for
brain tumor classification. International Journal of
Computer Science Issues (IJCSI), 9(3), 354.
[6] Bhima, K., &Jagan, A. (2016, April). Analysis of MRI
based brain tumor identification using segmentation
technique. In Communication and Signal Processing
(ICCSP), 2016 International Conference on (pp. 2109-
2113). IEEE.
[7] Vrji, K. A., &Jayakumari, J. (2011, July). Automatic
detection of brain tumor based on magnetic resonance
image using CAD System with watershed segmentation.
In Signal Processing, Communication, Computing and
Networking Technologies (ICSCCN), 2011 International
Conference on (pp. 145-150). IEEE.
[8] Rashid, M. H. O., Mamun, M. A., Hossain, M. A., &Uddin,
M. P. (2018, February). Brain Tumor Detection Using
Anisotropic Filtering, SVM Classifier and Morphological
Operation from MR Images. In 2018 International
Conference on Computer, Communication, Chemical,
Material and Electronic Engineering (IC4ME2)(pp.1-4).
IEEE.
[9] Sudharani, K., Sarma, T. C., &Rasad, K. S. (2015,
December). Intelligent Brain Tumor lesionclassification
and identification from MRI images using k-NN
technique. In Control, Instrumentation, Communication
and Computational Technologies (ICCICCT), 2015
International Conference on (pp. 777-780). IEEE.
[10] Vidyarthi, A., & Mittal, N. (2013, December). A hybrid
model for extraction of brain tumor in MR images. In
Signal Processing and Communication (ICSC), 2013
International Conference on (pp. 202-206). IEEE.
[11] Li, W., Chen, C., Su, H., & Du, Q. (2015). Local Binary
Patterns and Extreme Learning Machine for
Hyperspectral Imagery Classification. IEEE Trans.
Geoscience and Remote Sensing, 53(7), 3681-3693.
[12] Dhanaseely, A. J., Himavathi, S., &Srinivasan, E. (2012).
Performance comparison of cascade and feed forward
neural network for face recognition system.
[13] Liu, Q., & Liu, X. (2013, December). FeatureExtractionof
Human VirusesMicroscopic ImagesUsingGrayLevel Co-
occurrence Matrix. In Computer Sciences and
Applications (CSA), 2013 International Conference on
(pp. 619-622). IEEE.
[14] Singh, A. (2015, February). Detection of brain tumor in
MRI images, using combination of fuzzy c-means and
SVM. In Signal Processing and Integrated Networks
(SPIN), 2015 2ndInternationalConferenceon (pp. 98-
102). IEEE.

IRJET- Image Processing for Brain Tumor Segmentation and Classification

More Related Content

What's hot (20)

Similar to IRJET- Image Processing for Brain Tumor Segmentation and Classification (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Image Processing for Brain Tumor Segmentation and Classification