Brain Tumor Detection Using Deep-Learning Framework

Submitted in partial fulfillment of the requirements for

the award of
Bachelor of Engineering degree in Computer Science and
Engineering

by

J HARSHINI (Reg. No.

37110299) K GAYATRI (Reg. No.
37110303)

DEPARTMENT OF COMPUTER SCIENCE AND

ENGINEERING SCHOOL OF COMPUTING

SATHYABAMA
INSTITUTE OF SCIENCE AND TECHNOLOGY
(DEEMED TO BE UNIVERSITY)
Accredited with Grade “A” by NAAC
JEPPIAAR NAGAR, RAJIV GANDHI
SALAI, CHENNAI – 600 119

MARCH - 2021
 SATHYABAMA
INSTITUTE OF SCIENCE AND TECHNOLOGY
(DEEMED TO BE UNIVERSITY)
Accredited with “A” grade by NAAC
Jeppiaar Nagar, Rajiv Gandhi Salai, Chennai – 600 119
www.sathyabama.ac.in

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

BONAFIDE CERTIFICATE

This is to certify that this project report is the Bonafide work of JUJJAVARAPU
HARSHINI (Reg. No. 37110299) and KADAMUTHURU GAYATRI (Reg. No.
37110303) who carried out the project entitled “Brain Tumor Detection Using
Deeop-Learning Framework” under my supervision from August 2020 to March
2021.

Internal Guide
Dr. P. Asha, M.E.,Ph.D.,

Head of the Department

Dr. L. Lakshmanan, M.E., Ph.D.,
Dr. S. Vigneshwari, M.E., Ph.D.,

Submitted for Viva voce Examination held on

Internal Examiner External Examiner
 DECLARATION

I JUJJAVARAPU HARSHINI here by declare that the Project Report entitled

“Brain Tumor Detection Using Deep-Learning Framework” is done by me
under the guidance of Dr.P.ASHA, M.E.,Ph.d., Department of Computer Science
and Engineering at Sathyabama Institute of Science and Technology is submitted
in partial fulfillment of the requirements for the award of Bachelor of Engineering
degree in Computer Science and Engineering.

DATE:

PLACE: CHENNAI SIGNATURE OF THE CANDIDATE

 ACKNOWLEDGEMENT

I am pleased to acknowledge my sincere thanks to Board of Management of

SATHYABAMA for their kind encouragement in doing this project and for
completing it successfully. I am grateful to them.

I convey my thanks to Dr. T. Sasikala, M.E., Ph.D., Dean, School of Computing,

Dr. S. Vigneswari, M.E., Ph.D., and Dr. L. Lakshmanan, M.E., Ph.D., Heads of
the Department of Computer Science and Engineering for providing me
necessary support and details at the right time during the progressive reviews.

I would like to express my sincere and deep sense of gratitude to my Project

Guide D.P.Asha, M.E.,Ph.d., for her valuable guidance, suggestions and
constant encouragement paved way for the successful completion of my project
work.

I wish to express my thanks to all Teaching and Non-teaching staff members of

the Department of Computer Science and Engineering who were helpful in
many ways for the completion of the project.
 ABSTRACT

Using magnetic resonance imaging to manually segment brain tumors for cancer diagnosis is a

complex, tedious and time-consuming task. The accuracy and the robustness of brain tumor

segmentation, therefore, are crucial for the diagnosis, treatment planning, and treatment outcome

evaluation. Most automated brain tumor segmentation techniques use manually developed functions.

Traditional deep learning methods (such as convolutional neural networks) also require a large amount

of annotated data for training, which is usually difficult to obtain in the medical field. Here, we

describe a new model two-pathway-group CNN architecture for brain tumor segmentation, which

exploits local features and global contextual features simultaneously. The model uses the equivalence

of the bidirectional CNN model to reduce instability and overfit common parameters. Finally, we

merge the cascaded architecture into a two-way multicast CNN, where the output of the basic CNN is

processed as an auxiliary source and summarized at the final level. Validation of the models in the

data sets BRATS2013 and BRATS2015 shows that the integration of this group CNN into a pathway

architecture improved the overall performance over the currently published state-of-the-art while

computational complexity remains attractive.

 TABLE OF CONTENTS

ABSTRACT v
LIST OF FIGURES vii

CHAPTER No. TITLE PAGE No.

1. INTRODUCTION
1.1 Introduction 1

2. LITRECTURE SURVEY
3
3. AIM AND SCOPE
3.1 Aim 5
3.2 Synopsis 5

4. SYSTEM DESIGN AND METHODOLOGY

4.1 Existing System 6
4.2 Proposed System 6
4.3 Requirement Specification 6
4.4 Hardware and Software Specification 8
4.5 Software Description 8
4.6 Methodology 9
5. Design And Implementation
5.1 Nonfunctional Requirements 12
5.2 Sequence Diagram 14
5.3 use Case Diagram 15
5.4 Activity Diagram 16
5.5 Collaboration Diagram 17
5.6 Modules 18
5.7 Test Procedure 20
6. RESULT AND CONCLUSION 23

SCREEN SHOTS 25
SOURCE CODE 27
GENERAL PAPER 35
 LIST OF FIGURES

FIGURE No. FIGURE NAME PAGE No.

5.1 Architecture Diagram 13

5.2 Sequence Diagram 14
5.3 Use Case Diagram 15
5.4 Activity Diagram 16
5.5 Collaboration Diagram 17
E.1 Meningioma Tumor Classifier 25
E.2 Glioma Tumor Classifier 25
E.3 No Tumor Classifier 26
E.4 Pituitary Tumor Classifier 26
 CHAPTER 1

INTRODUCTI

ON

1.1 INTRODUCTION

Tumors will have a huge effect on the brain. Brain cells are destroyed in the area affected by tumor
gets and can cause brain collapse. The result of the tumor depends on the size and area affected in
the brain. The Brain connected each and every part of the body together to make it perfect sense. If
anything happens to the brain our whole system collapses. Some neurons in brain do not have
capability to regenerate and there some neurons which stops regeneration as a person age. If the
tumor is situated in any of that non- regenerative areas, a person might even lose one of his/her
senses. Discovering the tumor at an early stage can save a person’s life. Artificial Intelligence is
revolutionizing Healthcare in many areas such as Disease Diagnosis with medical imaging, Surgical
Robot, maximizing hospital efficiency. Deep learning has been proven to be superior in detecting
diseases from X-rays, MRI scans and CT scans which could significantly improve the speed and
accuracy of diagnosis. Tumors are located and diagnosed through a very keen medical procedure.
Magnetic Resonance Image (MRI) is one such process. We are going to train and validate our model
on MRI. These images are sent into to model to train it to detect and locate brain tumor.

Segmenting brain tumors in multi-modal imaging data is a challenging problem due to unpredictable
shapes and sizes of tumors. Deep Neural Networks (DNNs) have already been applied to
segmentation problems and have shown significant performance improvement compared to the
previous methods [4]. We use Convolutional Neural Networks (CNNs) to perform the brain tumor
segmentation task on the large dataset of brain tumor MR scans provided by BRATS2015. CNNs are
DNNs in which trainable filters and local neighborhood pooling operations are applied alternatingly
on the raw input images, resulting in a hierarchy of increasingly complex features. Specifically, we
used multi-modality information from T1, T1c, T2 and Flair images as inputs to different CNNs. The
multiple intermediate layers apply convolution, pooling, normalization, and other operations to
capture the highly nonlinear mappings between inputs and outputs. We take the output of the last
hidden layer of each CNN as the
representation of a pixel in that modality and concatenate the representations of all the modalities as
features to train a random forest classifier.

Magnetic resonance imaging (MRI) is widely used medical technology for diagnosis of various
tissue abnormalities, detection of tumors. The active development in the computerized medical
image segmentation has played a vital role in scientific research. This helps the doctors to take
necessary treatment in an easy manner with fast decision making. Brain tumor segmentation is a hot
point in the research field of Information technology with biomedical engineering. The brain tumor
segmentation is motivated by assessing tumor growth, treatment responses, computer-based surgery,
treatment of radiation therapy, and developing tumor growth models. Therefore, computer-aided
diagnostic system is meaningful in medical treatments to reducing the workload of doctors and
giving the accurate results. This chapter explains the causes, awareness of brain tumor segmentation
and its classification, MRI scanning process and its operation, brain tumor classifications, and
different segmentation methodologies.
 CHAPTER 2

LITRECTURE

SURVEY

Habib [1], used artificial convolutional neural network (ANN) to detect tumor with a similar brain
tumor dataset employed in this paper. He achieved 88.7% accuracy while testing. He used different
neural network which provided him better accuracy. The neural network consists of two max pooling
layers and one convolutional 2d (Convo2d) layer in sequential pattern.

Lin and Chang [2], used K-means clustering algorithms with color-based segmentation to track
objects of the brain tumor. K-means clustering clusters the similar places together with color. The
interesting part in this paper is, they clustered color-spaced image from greyscale with K-means
algorithm.

The researchers in [4] used MRI or second resonance images to detect brain tumors. They classified
MRI images which is complex because of the variation in brain tumor size and shapes. Decision
Tree classifier and Multi-Layer perceptron are the two supervised learning techniques to detect brain
tumor.

In this paper [7], the authors explained how misdiagnosis made by image processing or machine
learning affect us and showed that they never always give accurate solution or result. There is other
variable that need to be taken care while detecting brain tumor. In this paper, they used MRI
augmentation technique. Which is sending images in model from various angles and different
perspectives. This technique allows model to train on various new images and that got a good result
and scores. They used a CNN along with Link Net architecture.

Sharma and Komal [7] proposed a method that involves features detection of brain tumor and
classified based on MRI data. They used various filters and image segmentation over images.
Sinthia and Malathi [8] proposed a CNN that automated detection and segmentation of brain tumor.
Neural network was constructed using TensorFlow library. BRATS2015 dataset is used by them.
Guotai Wang [9], these researchers proposed an application of deep learning that can interact with
user. The application consists of different convolutional neural network 4 architectures that can
create a model and can make segmentation of MRI images and can even highlight certain brain
oragans. The interesting part is that user can tune and make these organs to pop up. They provided
multiple architectures and compared them to get perfect result.

In this paper the authors [10], proposed the essential task carried out by MRI images in diseases
diagnosis. They even shared few methods on MRI image processing to implement them in other
papers.

Google Net architecture codenamed Inception-v1 is the improved utilization of computing resources
inside the network [14]. The network with the inception architecture is faster than the network with
non- inception architecture. The Google Net architecture including the inception module uses
rectified linear activation function, average pooling layer and not fully connected layer and dropout
after removing fully connected layer.

Alex Net [15] architecture is deeper and much greater than Lent architecture. It consists of eight
layers, five convolutional layers most of them are followed by max pooling and three fully
connected layers. The output is the 1000-way SoftMax that represents the classes. It is trained on
two parallel GTX 580 GPU 3 GB which communicate only in certain layers. This scheme reduces
the top-5 error rates. Alex Net is improved with Zeit architecture which visualizes the Alex Net
activities within the layers to debug problems and obtain better results. It allows observing the
evolution of features during training and maps the activities back to the pixel space in intermediate
layers.
 CHAPTER 3

Aim and Scope

3.1 Aim:

To detected the Brain Tumor using Two-Pathway-Group Conventional Neural Networks. Use the
traditional two-way cluster neural network for brain tumor detection.

3.2 Synopsis:

Brain tumor identification is really challenging task in early stages of life. But now it became

advanced with deep-learning. Now a day’s issue of brain tumor automatic identification is of

great interest. In Order to detect the brain tumor of a patient we consider the data of patients

like MRI images of a patient’s brain. Here our problem is to identify whether tumor is present

in patients’ brain or not. It is very important to detect the tumors at starting level for a healthy

life of a patient. There are many literatures on detecting these kinds of brain tumors and

improving the detection accuracies. In this paper, we estimate the brain tumor severity using

Convolutional Neural Network algorithm which gives us accurate results.

 Chapter 4

SYSTEM DESIGN &

METHODOLOGY

4.1 EXISTING SYSTEM

Existing systems describe the automation of cell segmentation. The technique

is used to interactive multi label segmentation for N dimensional images. It segments the

areas which are more difficult to segment. This method is iterative and provides feedback to

the user as the segment is calculated.

4.2 PROPOSED SYSTEM

We take a second dimension to propose a new strategy for MRI of the patient's brain. Here
preprocessing is done with Gaussian, which can be a line filter. Then, by identifying the area of the
tumor, the GLCM functions are used to extract functions from the image. CNN architecture of
bidirectional clusters for brain tumor segmentation, At the same time, it uses its own functions and
international information dissemination functions. This model provides equivalence in the
two-channel CNN model to reduce backward jitter and overcome parameter sharing. Finally, we
integrated the cascaded architecture into the dual-channel CNN pool, during which the basic CNN
pins were processed as additional sources, and finally, combined the CNN into a two-way
architecture, thereby improving the overall performance. Compared with the currently disclosed
progressive method, this method is improved, and the complexity of the process is still quite
pleasant.

4.3 REQUIREMENT SPECIFICATIONS

CANCEROUS brain tumors present themselves as unnatural, uncontrolled growth and division of
cells in the brain. Although brain tumors are not very common, they are one of the deadliest cancers.
For example, in the United States alone, approximately 23,000 new cases of brain cancer were
diagnosed in 2015. It is an abnormality in the brain tissues that damage the nervous system severely,
which result patient death. are the most common brain tumors that are infiltrative in nature, and
occur near white matter fibers. They may spread to any part of the brain making it difficult to detect.
They are considered
high-grade gliomas.
one of the most aggressive tumors with a median survival of 15 months. Glioma can be measured by
MRI using a variety of sequences, such as: B. T2-weighted reversal recovery (transition) with fluid
attenuation, T1-weighted(T1), contrast-enhanced T1 (T1c) and T2-weighted (T2), use existing
automated techniques to segment brain tumors. Healthy brains consist of three types of tissues: gray
matter, white matter and cerebrospinal fluid. Detection and segmentation of cancerous cells using
MRI not only helps to detect the presence of tumors and their location, but it also enables the
identification of tumor size, necrotic tissue, tumorous tissue (vascularized or not) and edema
(swelling near the tumor). Brain tumors vary in shape and appearance (gliomas look the same as
gliomas, strokes, etc.), which makes them difficult to separate radiologists. Furthermore, they may
appear at any location in the brain: depending on the origin of the brain tumor, they can be classified
as either primary
tumors or metastatic brain tumors. The edges of brain tumors are often ambiguous and fuzzy, and
are hard to distinguish from healthy tissues. Therefore, a more sensitive alternative to MRI is needed
to improve the detection of tumors and to increase the survival rate of people with brain

tumors. Machine-assisted image segmentation and subsequent quantification of cancer tissue

provide valuable information for early diagnosis and characterization of neuropathology, which can
then be used for appropriate treatment. Anatomical structure. In addition, this is very important for
early diagnosis, and it helps early prevention by formulating treatment strategies. Cancerous cells
are normally quantified by means of the number of lesions, their volume, and biomarkers that have
been shown to be related to cognitive deficits. As a result, the quantitative analysis of effected
regions requires accurate lesion segmentation, this is challenging due to differences in the size,
shape, location and frequency of cancer lesions. Perhaps the most accurate brain tumor segmentation
results are achieved manually by an expert; however, this is an expensive, time-consuming, tedious,
An impractical task, error-prone, and dependent on differences between observers. As a result,
doctors usually only use qualitative or visual inspections, or at best only rough measurements, such
as the approximate volume and number of tumors. Manual segmentation of brain tumors from large
MRI images is a difficult and time-consuming task. The existing methods of segmenting brain
tumors can usually be divided into generative models or discriminatory models [1]. Generative
models require prior information and segmentation of brain tumors, whereas discriminative models
depend on a set of features and classifiers. The most commonly adopted classifiers are support
vector machines (SVMs), random forests, neural
genetic algorithm and network. In contrast, automatic brain tumor segmentation methods use hand-
designed features and a variety of image features (e.g., shape, area, perimeter, circularity etc.),
intensity (e.g., mean, variance, standard deviations) and texture (e.g., contrast, entropy, correlation
etc.).
Recently, deep learning, and in particular the Convolutional Neural Network

(CNN) has become the methodology of choice for medical image analysis, following its tremendous
success in routine computer vision applications [2], [3]. Regarding the detection of tumors, the
generation of candidates and the reduction of false positives, using deep learning-based methods,
unambiguously outperformed traditional machine learning approaches [4]– [7]. This achievement
was acknowledged in 2015 where the Deep Medic software for brain lesion segmentation, which
was based on a 3D-CNN coupled with a 3D fully connected CRF, won the ISLES 2015 challenge
[8]. Additional deep learning-based brain tumor segmentation methods were presented in the 2013,
2015 and 2017 challenges. Different deep learning models were adopted, including FCNN [9]– [4],
[11], 3DCNN [12],
[13], FCNN with CRF [14], 3D U-Net [15], [16] and Autoencoders [17], [18].

4.4 HARDWARE AND SOFTWARE SPECIFICATION

4.4.1 HARDWARE REQUIREMENTS

● Hard Disk : 500GB and Above

● RAM : 4GB and Above

● Processor : I3 and Above

4.4.2 SOFTWARE REQUIREMENTS

✔ Operating System : Windows 7, 8, 10 (64 bit)

✔ Software : Python

✔ Tools : Anaconda (Jupyter Note Book IDE)

4.5 SOFTWARE DESCRIPTION:

PYTHON
• Python is a free, open-source programming language. Therefore, all you have to do is
install Python once, and you can start working with it. Not to mention that you can contribute
your own code to the community. Python is also a cross-platform compatible language. So, what
does this mean? Well, you can install and run Python on several operating systems.
• Python is also a great visualization tool. It provides libraries such as Matplotlib, seaborn and
bokeh to create stunning visualizations.
B. PANDAS
• Pandas is a popular Python package for data science, and with good reason: it offers
powerful, expressive and flexible data structures that make data manipulation and analysis easy,
among many other things. The Data Frame is one of these structures. Pandas is a highlevel data
manipulation tool
developed by Wes McKinney. It is built on the Numpy package and its key data structure is called the
Data Frame. Data Frames allow you to store and manipulate tabular data in rows of observations and
columns of variables.
• Pandas is built on top of the NumPy package, meaning a lot of the structure of NumPy is used
or replicated in Pandas. Data in pandas is often used to feed statistical analysis in SciPy, plotting
functions from Matplotlib, and machine learning algorithms in Scikit-learn. There are two types of
data structures in pandas: Series and Data Frames.
a) Series: A pandas Series is a one-dimensional data structure
b) Data Frame: A pandas Data Frame is a two (or more) dimensional data structure basically
a table with rows and columns.

4.6 METHODOLOGY
4.6.1 CONVOLUTIONAL NEURAL NETWORK
A Convolutional Neural Network (ConvNet/CNN) is a Deep Learning algorithm which can take in
an input image, assign importance (learnable weights and biases) to various aspects/objects in the
image and be able to differentiate one from the other. The pre-processing required in a ConvNet is
much lower as compared to other classification algorithms. While in primitive methods filters are
handengineered, with enough training, ConvNets have the ability to learn these
filters/characteristics. The architecture of a ConvNet is analogous to that of the connectivity pattern
of Neurons in the Human Brain and was inspired by the organization of the Visual Cortex.
Individual neurons respond to stimuli only in a restricted region of the visual field known as the
Receptive Field. A collection of such fields overlaps to cd over the entire visual area. The objective
of the Convolution Operation is to extract the high-level features such as edges, from the input
image. ConvNets need not be limited to only one Convolutional Layer. Conventionally, the first
Cavalier is responsible for capturing the LowLevel features such as edges, color, gradient
orientation, etc. With added layers, the architecture adapts to the High-Level features as well, giving
us a network, which has the wholesome understanding of images in the dataset, similar to how we
would.

4.6.2 MAGNETIC RESONANCE IMAGING (MRI)

The MRI is a diagnostic tool used for analyzing and studying the human anatomy. The medical
images acquired in various bands of the electromagnetic spectrum. The wide variety of sensors used
for the acquisition of images and the physics behind them, make each modality suitable for a
specific purpose. In MRI, the pictures are produced using a magnetic field, which is approximately
10,000 times stronger than the earth’s magnetic field. The MRI produces more detailed images than
other techniques, such as CT or ultrasound. The MRI also provides maps of anatomical structures
with a high soft-tissue contrast. Basically, the magnetic resonance of hydrogen (1H) nuclei in water
and lipid is measured by an MRI
scanner. As the signal values are 12-bit coded, 4096 shades can be represented by a pixel [11]. The
MRI scanners require a magnetic field and it is available at 1.5 or 3 T. In comparison with the
earth’s magnetic field (~50 ft.) the magnetic field of a 3 T MRI scanner is approximately 60,000
times the earth field. The patient is placed in a strong magnetic field, which causes the protons in the
water molecules of the body to align either in a parallel or anti-parallel orientation with the magnetic
field. A radiofrequency pulse is introduced, causing the spinning protons to move out of the
alignment. When the pulse is stopped, the protons realign and emit radio frequency energy signal
that is localized by the magnetic fields and are spatially varied and rapidly turned on and off. A
radio antenna within the scanner detects the signal and creates the image. Functional MRI is a
technique for examining the brain activation, which unlike PET, is non-invasive with relatively high
spatial resolution.

Figure 4.6 MRI of Human Brain

The most common method utilizes a technique called blood oxygen level dependent contrast. This is
an example of endogenous contrast, making use of the inherent signal differences in blood
oxygenation content. In the normal resting state, a high concentration of deoxyhemoglobin
attenuates the MRI signal due to its paramagnetic nature. However, the neuronal activity, in response
to some task or stimulus, creates a local demand for the oxygen supply, which increases the fraction
of oxy hemoglobin causing a signal increase on T2 or T2*-weighted images. In a typical experiment,
the patient is subjected to a series of rest and task intervals, during which MRI images are repeatedly
acquired. A radio antenna within the scanner detects the signal and creates the image. Functional
MRI is a technique for examining the brain activation, which unlike PET, is non-invasive with
relatively high spatial resolution. The signal changes during the course of time are then examined on
a pixelby-pixel basis to test how well they correlate with the known stimulus pattern. The pixels that
demonstrate a statistically significant correlation are highlighted in color and overlaid onto a
grayscale MRI image to create an activation map of the brain. The location and extent of activation
is linked to the type of stimulus. Thus, a simple thumbfinger movement task will produce activation
in the primary motor cortex.
 Chapter 5
DESIGN AND IMPLEMENTATION

Constraints in Analysis

♦ Constraints as Informal Text

♦ Constraints as Operational Restrictions

♦ Constraints Integrated in Existing Model Concepts

♦ Constraints as a Separate Concept

♦ Constraints Implied by the Model Structure

Constraints in Design

♦ Determination of the Involved Classes

♦ Determination of the Involved Objects

♦ Determination of the Involved Actions

♦ Determination of the Require Clauses

♦ Global actions and Constraint Realization

Constraints in Implementation

The hierarchy of relationships can lead to more classes and more complex
implementation structures. Therefore, it is recommended to convert the hierarchical structure of the
relationship to a simpler structure, such as a classic plane structure. The developed hierarchical model
can be easily converted into a flat two-part model. On the one hand it is composed of classes, on the
other it is composed of flat relationships. Since flat relationships are easy to implement, they are
preferred at the design level. There are no logos or functions related to the plane relationship. The
plane relationship follows the concept of entity relationship modeling and many object-oriented
technologies.
5.1 Nonfunctional Requirements
Performance Requirement: The application at this side controls and communicates with the
following three main general components. embedded browser in charge of the navigation and
accessing to the web service.

Server Tier: The server side contains the main parts of the functionality of the proposed
architecture. The components at this tier are the following.
Web Server, Security Module, Server-Side Capturing Engine, Preprocessing Engine, Database
System, Verification Engine, Output Module.
Safety Requirements
1.Software is essential for safety. In this case, there are issues related to the integrity level.
2. Even if the software is part of a security-critical system, it may not necessarily be critical to
security. For example, a program can only write transactions.
3. If the system is to have a higher integrity level, and it is shown that the software has this
integrity level, the hardware must have at least the same integrity level.
4. If the hardware and software of the system are (in a broad sense) unreliable, it makes no
sense to write "perfect" code in any language.
5. If a computer system is to run software with a higher integrity level, the system should
not support software with a lower integrity level at the same time.
6.Systems with different security level requirements should be separated.
7.Otherwise, the highest level of integrity should be applied to all systems in the same environment.
Architecture Diagram:

Fig: 5.1
 5.2 Sequence Diagram:

A sequence diagram is an interaction diagram that shows how and in what order the
processes interact. This is the construction of message sequence diagrams, sometimes called
event diagrams, event scenarios, and sequence diagrams.

Fig: 5.2
5.3 Use Case Diagram:

Unified Modeling Language (UML) is a universal, standardized modeling language used for
software development. This standard is managed and created by the asset management team. UML
contains many graphical markup techniques for creating visual models of software-intensive
systems. It is used to define, visualize, modify, construct and record the artifacts of object-oriented,
software-intensive systems under development. Unified Modeling Language (UML) is a universal
modeling language that has been standardized in the field of software development. This standard is
managed and created by the asset management team. It is used to define, visualize, modify,
construct and document the artifacts of intensively developed object-oriented software systems.
Use case Diagram
Use case diagrams are used to graphically describe the functions provided by the system based on
participants, their goals, and any dependencies between these use cases. The use case diagram
consists of two parts:
Use case: A use case describes a serie s of actions through which the subject can be measured
and drawn as a horizontal ellipse.
Participants: Participants are individuals, organizations, or external systems that play a role in
one or more interactions with the system.

Fig: 5.3
5.4 Activity Diagram:

Activity diagram is a graphical representation of workflows of stepwise activities and

actions with support for choice, iteration and concurrency. An activity diagram shows the
overall flow of control.

The most important shape types:

● Rounded rectangles represent activities.

● Diamonds represent decisions.
● Bars represent the start or end of concurrent activities.
● A black circle represents the start of the workflow.
● An encircled circle represents the end of the workflow.

Fig: 5.4
 5.5 Collaboration Diagram:

UML collaboration diagram illustrates the relationship and interaction between software objects. They
assume that use cases, system usage contracts and domain models already exist. The collaboration
diagram shows the messages sent between the class and the object.

Fig: 5.5
5.6 MODULES

❖ Image acquisition
❖ Image preprocessing
❖ Image segmentation
❖ Convolutional neural network
❖ Tumor detection

MODULE EXPLANATION:

5.6.1 Image Acquisition

The Primary Phase is acquiring images. After the Images collection, the obtained images have to be
prepared with a wide range of vision. First capture the input images from available source

5.6.2 Pre-Processing

The images which are collected are subjected to pre- processing. In Pre- processing stage basic steps
are image resizing and applying Gaussian filters for a perfect input clear image for easy
identification of an image. Pre-processed images will be segmented digitally into various pixels. We
do this segmentation for an image is to modify its representation to have more clarity to analyze the
images.

5.6.3 image segmentation

In the first stage, the pre-processed brain Magnetic Resonance image will be transformed into a
binary
image with a threshold of 128 for the cutoff. Pixel values higher than the specified thresholds are
mapped as white, with other regions marked as black; these two allow various regions to be
generated around the disease. In the second stage, an erosion process of morphology is used to
extract white pixels. Eventually, the eroded area and the original image are separated into two
equal areas, and the region with black pixels from the eroding is counted as a mask of brain
Magnetic Resonance image. In this paper, wavelet transformation is used for the efficient
segmentation of the brain Magnetic Resonance image. Figure 3 shows the fully automatic
heterogeneous segmentation. Figure 3(a) shows the axial image and its segmentation figure 3(b)
Coronal image and its segmentation figure 3 (c) Sagittal images and its segmentation.

Featur
e
Extraction

In the feature extraction process, we can implement the effective texture operator

which labels the pixels of an image. Here we extract the features and characteristics of

Images for easy detection of brain tumor.

Classification

Convolutional neural networks algorithm is used for classification of brain images. It

is producing the best results for the image

 Tumor Detection

Finally, analyze the image using filters and Convolutional neural networks algorithm to

detect the tumor or Non-tumor.

5.7 TEST PROCEDURE

System Testing:
Testing is performed to identify errors. It is used for quality assurance.

Testing is an integral part of the entire development and maintenance process. The goal of the

testing during phase is to verify that the specification has been accurately and completely

incorporated into the design, as well as to ensure the correctness of the design itself. For

example, the design must not have any logic faults in the design is detected before coding

commences, otherwise the cost of fixing the faults will be considerably higher as reflected.

Detection of design faults can be achieved by means of inspection as well as walkthrough.

Testing is one of the important steps in the software development phase. Testing checks

for the errors, as a whole of the project testing involves the following test cases:

⮚ Static analysis is used to investigate the structural properties of the Source code.

⮚ Dynamic testing is used to investigate the behavior of the source

code by executing the program on the test data.

5.7.1. TEST DATA AND

OUTPUT UNIT TESTING

Unit testing is conducted to verify the functional performance of each modular

component of the software. Unit testing focuses on the smallest unit of the software design (i.e.),
 the module. The white-box testing techniques were heavily employed for unit testing.

5.7.2 FUNCTIONAL TESTS

Functional test cases involved exercising the code with nominal input values

for which the expected results are known, as well as boundary values and special values, such

as logically related inputs, files of identical elements, and empty files.

Three types of tests in Functional test:

⮚ Performance Test

⮚ Stress Test

⮚ Structure Test

PERFORMANCE TEST

It determines the amount of execution time spent in various parts of the unit,

program throughput, and response time and device utilization by the program unit.

STRESS TEST

Stress Test is those tests designed to intentionally break the unit. A Great deal can

be learned about the strength and limitations of a program by examining the manner in which

a programmer in which a program unit breaks.

STRUCTURED TEST

Structure Tests are concerned with exercising the internal logic of a program and

traversing particular execution paths. The way in which White-Box test strategy was

employed to ensure that the test cases could Guarantee that all independent paths within a

module have been

 have been exercised at least once.

⮚ Exercise all logical decisions on their true or false sides.

⮚ Execute all loops at their boundaries and within their operational bounds.

⮚ Exercise internal data structures to assure their validity.

⮚ Checking attributes for their correctness.

⮚ Handling end of file condition, I/O errors, buffer problems and textual

errors in output information

Integration testing is a systematic technique for construction the program structure while at the

same time conducting tests to uncover errors associated with interfacing. i.e., integration testing

is the complete testing of the set of modules which makes up the product. The objective is to

take untested modules and build a program structure tester should identify critical modules.

Critical modules should be tested as early as possible. One approach is to wait until all the units

have passed testing, and then combine them and then tested. This approach is evolved from

unstructured testing of small programs. Another strategy is to construct the product in

increments of tested units. A small set of modules are integrated together and tested, to which

another module is added and tested in combination. And so on. The advantages of this approach

are that, interface dispenses can be easily found andcorrected.

The major error that was faced during the project is linking error. When all the

modules are combined the link is not set properly with all support files. Then we checked out

for interconnection and the links. Errors are localized to the new module and its

intercommunications. The product development can be staged, and modules integrated in as

they complete unit testing.

Testing is completed when the last module is integrated and tested.

 CHAPTER 6

RESULT AND CONCLUSION

Result:

Our data set contains tumor and non-tumor MRI images obtained from various online sources. Use

convolutional neural network for detection. Modeling is done using Python language. Calculate the

accuracy and compare it with all other modern methods.

To determine the effectiveness of the proposed brain, training accuracy, verification accuracy, and

verification loss need to be calculated. Tumor classification scheme. The current technology for

detecting brain tumors uses SVM (Support Vector Machine) classification. Feature extraction

requires output. Based on the feature value, the classification output is generated and the

accuracy is calculated. Tumor and non-tumor detection based on support vector machines take a

long time and have poor calculation accuracy. The proposed CNN-based classification does not

require a separate feature extraction step. The value of this function is taken from CNN itself. In

the picture. The classification results of tumor and non-tumor brain imaging are shown. Therefore,

the complexity and calculation time are low and accurate. The figure shows the results of brain

tumor classification accuracy. Finally, according to the value of the probability score, it is classified

as brain tumor or non- tumor brain. Normal brain imaging is the least likely. The score value

compared with normal and

neoplastic brains.

CONCLUSION

Our data set includes tumor MRI images and non-tumor images obtained from various online

sources. Radiation podia contains real patient cases. Tumor images are obtained from the test

data set of "Radio podia and Brain Tumor Image Segmentation Benchmark (BRATS) 2015". The

detection is carried out through a convolutional network. Modeling is done using Python language.

Calculate the accuracy and compare it with all other modern methods. In order to determine the

effectiveness of the proposed brain, training accuracy, verification accuracy, and verification loss

need to be calculated.

Tumor classification scheme. The current technology for detecting brain tumors uses SVM (Support

Vector Machine) classification. Feature extraction requires output. Based on the feature value, the

classification output is generated and the accuracy is calculated. Tumor and non-tumor detection

based on support vector machines take a long time and have poor calculation accuracy. The

proposed CNN-based classification does not require a separate feature extraction step. The value

of this function is taken from CNN itself. In the picture. The classification results of tumor and

non-tumor brain imaging are shown. Therefore, the complexity and calculation time are low and

accurate. The figure shows the results of brain tumor classification accuracy. Finally, according to

the value of the probability score, it is classified as brain tumor or non-tumor brain. Normal brain

imaging is the least likely. The score value compared with normal and neoplastic brains.
SCREEN SHOT

Figure : E.1 Meningioma Tumor classifier

Figure: E.2 Glioma Tumor Classifier

 Figure: E3 No Tumor Classifier

Figure: E4 Pituitary Tumor Classifier

 SOURCE CODE

Import libraries

Nc import numpy as
np import pandas as
pd import seaborn
as sns
import matplotlib.pyplot as
plt import cv2
from sklearn.model_selection
import train_test_split
Store every image in an array
import os
data_dir = ('D:/Brain-Tumor-Classification/Training/')
categories = ['glioma_tumor', 'meningioma_tumor', 'no_tumor',
'pituitary_tumor'] for i in categories:
path =
os.path.join(data_dir, i) for
img in os.listdir(path):
img_array = cv2.imread(os.path.join(path,img))

Resize each image to same size for fast processing

img_size = 128
image_array = cv2.resize(img_array, (img_size,img_size))
gt1 = cv2.imread('D:/Brain-Tumor-Classification/Training/glioma_tumor/gg(1).jpg')
mt1 =
cv2.imread('D:/Brain-Tumor-Classification/Training/meningioma_tumor/m
(10).jpg')nt1 = cv2.imread('D:/Brain-Tumor-Classification/Training/no_tumor/1.jpg') pt1 =
cv2.imread('D:/Brain-Tumor-Classification/Training/pituitary_tumor/p(151).jp
g')
Example of Glioma Tumor
plt.rcParams["figure.figsize"] =
(5,5) plt.imshow(gt1)
plt.axis('off')
Example of Meningioma Tumor
plt.rcParams["figure.figsize"] =
(5,5) plt.imshow(mt1)
plt.axis('off')
Example of No Tumor
plt.rcParams["figure.figsize"] =
(5,5) plt.imshow(nt1)

plt.axis('off')

Convert each image to grayscale and append into an array

train_data = []

for i in categories:
train_path =
os.path.join(data_dir,i) tag =
categories.index(i)
for img in
os.listdir(train_path): try:
image_arr = cv2.imread(os.path.join(train_path , img),
cv2.IMREAD_GRAYSCALE)
new_image_array = cv2.resize(image_arr, (img_size,img_size))
train_data.append([new_image_array , tag])
except Exception
as e: pass

Import keras' functions to create CNN

modelfrom sklearn.metrics import
confusion_matrix import itertools

from keras.utils.np_utils import to_categorical

 from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten, Conv2D,
MaxPool2D from keras.optimizers import RMSprop,Adam
from keras.preprocessing.image import
ImageDataGenerator from keras.callbacks import
ReduceLROnPlateau
from keras.callbacks import

EarlyStopping model = Sequential()

model.add(Conv2D(filters = 64, kernel_size = (5,5),padding = 'Same', activation ='relu',

input_shape = (128,128,1)))
model.add(MaxPool2D(pool_size=(2,2)))
model.add(Dropout(0.2))

model.add(Conv2D(filters = 128, kernel_size = (3,3),padding = 'Same', activation

='relu'))
model.add(MaxPool2D(pool_size=(2,2),
strides=(2,2))) model.add(Dropout(0.2))

model.add(Conv2D(filters = 128, kernel_size = (3,3),padding = 'Same', activation

='relu'))
model.add(MaxPool2D(pool_size=(2,2),
strides=(2,2))) model.add(Dropout(0.2))

model.add(Conv2D(filters = 128, kernel_size = (2,2),padding = 'Same', activation

='relu'))
model.add(MaxPool2D(pool_size=(2,2),
strides=(2,2))) model.add(Dropout(0.2))

model.add(Conv2D(filters = 256, kernel_size = (2,2),padding = 'Same', activation

='relu'))
model.add(MaxPool2D(pool_size=(2,2),
strides=(2,2))) model.add(Dropout(0.2))

model.add(Flatten())

model.add(Dense(1024, activation =
"relu")) model.add(Dropout(0.5))

model.add(Dense(4, activation = "softmax"))

optimizer = Adam(lr=0.001, beta_1=0.9, beta_2=0.999)

model.compile(optimizer = optimizer , loss = "categorical_crossentropy",

metrics=["accuracy"])

epochs = 20

es = EarlyStopping(
monitor='val_acc'
, mode='max',
patience = 3
)

batch_size = 16
imggen =
ImageDataGenerator(
featurewise_center=Fal
se,

samplewise_center=False,
featurewise_std_normalization=Fa
lse,
samplewise_std_normalization=Fa
lse, zca_whitening=False,
rotation_range=0,
 zoom_range = 0,
width_shift_range
=0,
height_shift_range
=0,
horizontal_flip=Tru
e,
vertical_flip=False)
Fit the model with Train and Validation data sets
from keras.callbacks import ModelCheckpoint

checkpoint = ModelCheckpoint("model_weights.h5", monitor='val_acc', verbose=1,

save_best_only=True, mode='max')
callbacks_list = [checkpoint]

imggen.fit(X_train)
history = model.fit_generator(imggen.flow(X_train,y_train,batch_size =
batch_size), epochs = epochs, validation_data =
(X_val,y_val), steps_per_epoch = X_train.shape[0] //
batch_size,
callbacks=callbacks_
list) serialize model structure to
JSON model_json =
model.to_json()
with open(r"D:\Brain-Tumor-Classification\model.json", "w") as
json_file: json_file.write(model_json
REFERENCE

[1]Abhishek Anil, Aditya Raj, H Aravind Sarma, Naveen Chandran R,Deepa

P L, “Brain Tumor detection from brain MRI using Deep Learning”

International

Journal of Innovative Research in Applied Sciences and Engineering

(IJIRASE), Volume 3, Issue 2, DOI:10.29027/IJIRASE.v3.i2.2019, 458-465,

August 2019.

[2] George, Dena Nadir, Hashem B. Jehlol, and Anwer Subhi

Abdulhussein. "Brain Tumour Detection Using Shape features and

Machine Learning

Algorithms".

[3] Işın, Ali, Cem Direkoğlu, and Melike Şah. "Review of MRI-based brain

tumor image segmentation using deep learning methods." Procedia

Computer

Science 102 (2016): 317-324

[4] Leiner, Tim, et al. "Machine learning in cardiovascular magnetic

resonance: basic concepts and applications." Journal of Cardiovascular

Magnetic

Resonance 21.1 (2019): 61.

[5] Lundervold, Alexander Selvikvåg, and Arvid Lundervold. "An overview of

deep learning in medical imaging focusing on MRI." Zeitschrift für

 Medizinische

Physik 29.2 (2019): 102-127.

[6] Long, C., Basharat, A., Hoogs, A.: A Coarse-to-fine Deep

Convolutional Neural Network Framework for Frame Duplication

Detection and Localization

 in Forged Videos, p. 10 (2018)

[7] M. Wu, C. Lin and C. Chang, "Brain Tumor Detection Using

ColorBased K- Means Clustering Segmentation," Third International

Conference on Intelligent

Information Hiding and Multimedia Signal Processing (IIH-MSP

2007), Kaohsiung, 2007.

[8] Mohsen, Heba, et al. "Classification using deep learning neural

networks for brain tumors." Future Computing and Informatics Journal

3.1 (2018): 68-71.

[9] Nogovitsyn, N., et al.: Testing a deep convolutional neural

network for automated hippocampus segmentation in a longitudinal

sample of healthy

participants. NeuroImage 197, 589–597 (2019).

[10] Sharma, Komal, Akwinder Kaur, and Shruti Gujral. "Brain tumor

detection based on machine learning algorithms." International Journal of

Computer

Applications 103.1 (2014).

[11] Sobhaninia, Zahra, et al. "Brain tumor segmentation using deep

learning by type specific sorting of images." arXiv preprint

arXiv:1809.07786 (2018).

[12] Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.:
Rethinking the Inception Architecture for Computer Vision,

arXiv:1512.00567 [cs], December

2015

[13]Wang, Guotai, et al. "Interactive medical image segmentation using

deep learning with image-specific fine tuning." IEEE transactions on

medical imaging

37.7 (2018): 1562-1573.

[14] Zacharaki, E.I., et al.: Classification of brain tumor type and grade

using MRI texture and shape in a machine learning scheme. Magn. Reson.

Med. 62(6),

1609–1618 (2009). https://doi.org/10.1002/mrm.22147

[15] https://en.wikipedia.org/wiki/Convoluti

onal_neural_network , December, 21

2019

[16] https://en.wikipedia.org/wiki/Loss_function , December, 21 2019

[17] https://en.wikipedia.org/wiki/Mathematical_optimization , December, 21 2019.

 GENERAL PAPER

A)
B)