Brain Tumor Detection Using MobileNet Architecture

CHAPTER 1
INTRODUCTION
Brain tumors are a significant health concern worldwide, often requiring timely diagnosis and
intervention to improve patient outcomes. The accurate and early detection of brain tumors is
crucial for effective treatment planning and prognosis. In recent years, deep learning techniques
have emerged as powerful tools for automated medical image analysis, offering the potential to
enhance the efficiency and accuracy of brain tumor detection.

Among various deep learning architectures, MobileNet has gained prominence due to its
efficiency in resource-constrained environments. Making it suitable for real-time analysis of
medical images using readily available hardware, such as smartphones and tablets. Leveraging
the capabilities of Mobile Net for brain tumor detection can potentially enable prompt and
accessible diagnoses, particularly in remote or low-resource healthcare settings.

In this Report, we present a comprehensive study on brain tumor detection using the Mobile Net
model. Our objective is to develop a system that combines the computational efficiency of
Mobile Net with the accuracy required for effective brain tumor identification. By employing
deep learning techniques, our research aims to address the challenges associated with manual
interpretation and the potential for human error in brain tumor detection.

1.1 HISTORY
Brain tumors are abnormal growths of cells in the brain that can cause various neurological
symptoms and potentially life-threatening conditions. Early detection and accurate diagnosis of
brain tumors are crucial for effective treatment planning and improving patient outcomes.
Traditional methods of brain tumor detection involve manual examination of MRI scans by
radiologists, which can be time-consuming and subject to human error.

In recent years, deep learning techniques have revolutionized the field of medical
imaging analysis, offering promising solutions for automated tumor detection and classification.

Department of CSE, JSSATEB 2024-2025 Page 1

Brain Tumor Detection Using MobileNet Architecture

Convolutional Neural Networks (CNNs), a type of deep learning model, have shown remarkable
performance in various image classification tasks, including medical image analysis.

1.2 PROBLEM STATEMENT

The detection of brain tumors is a critical task in the field of medical diagnostics.
Traditional methods of brain tumor detection rely on manual examination of MRI scans by
radiologists, which can be time-consuming and subject to human error. The need for automated
and efficient brain tumor detection systems has become increasingly important to improve
diagnostic accuracy, reduce analysis time, and enhance patient outcomes.
The existing challenges in brain tumor detection include:
• Manual Examination: The current practice of manually examining MRI scans for brain tumor
detection is labor-intensive and can lead to variations in interpretation among different
radiologists. Human error and subjectivity may result in delayed or inaccurate diagnoses,
potentially affecting patient treatment.
• Time Constraints: The manual examination process is time-consuming, especially when
dealing with many MRI scans. This delay can impede timely treatment planning for patients with
brain tumors, potentially leading to adverse health consequences.
• Limited Resources: In many medical settings, there is a shortage of expert radiologists,
particularly in regions with limited healthcare resources. Automating the brain tumor detection
process can alleviate the burden on radiologists, allowing them to focus on more critical tasks
and providing healthcare access to underserved areas.
• Complexity and Variability: Brain tumors exhibit diverse characteristics, including size,
location, and shape, making their detection and classification challenging. The complexity and
variability of brain tumor images require sophisticated techniques that can capture subtle patterns
and features indicative of tumors.

1.3 OBJECTIVES
The objectives of this study on brain tumor detection using the MobileNet architecture are as
follows:
• Develop an automated system: Design and implement a web application that utilizes deep
learning techniques and the MobileNet architecture for automated brain tumor detection in MRI

Department of CSE, JSSATEB 2024-2025 Page 2

Brain Tumor Detection Using MobileNet Architecture

scans. The application should be capable of accurately analyzing the scans and providing reliable
tumor detection results.
• Improve diagnostic accuracy: Enhance the accuracy of brain tumor detection by leveraging
the power of deep learning algorithms. Train the MobileNet model on a large dataset of MRI
scans to learn and recognize the specific features and patterns associated with brain tumors.
• Reduce analysis time: Reduce the time required for brain tumor analysis by automating the
detection process. The developed system should efficiently analyze MRI scans and provide
timely results, enabling medical professionals to make faster and more informed diagnostic
decisions.

Department of CSE, JSSATEB 2024-2025 Page 3

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 2
LITERATURE SURVEY
Dr. Chinta Someswararao [1], paper titled “Brain Tumor Detection Model from MR
Images using Convolutional Neural Network“was a combination of CNN model classification
problem for predicting whether the subject has brain tumor or not & Computer Vision problem
for automate the process of brain cropping from MRI scans. The final accuracy achieved by him
is much higher than 50% baseline (random guess). However, it could further be increased by
larger number of train images or through different models and techniques.

Masoumeh Siar and Mohammad Teshnehlab [2], paper titled “Brain Tumor
Detection Using Deep Neural Network and Machine Learning Algorithm” used the
combination of feature extraction algorithm and the CNN for tumor detection from brain images
is presented. The CNN is capable of detecting a tumor. The CNN is very useful for selecting an
auto-feature in medical images. Images collected at the centers were labeled by clinicians, then,
tumor screenings were categorized into two normal and patient classes.

Devkota et al. [3] established the whole segmentation process based on Mathematical
Morphological Operations and spatial FCM algorithm which improves the computation
time, but the proposed solution has not been tested up to the evaluation stage and outcomes as-
Detects cancer with 92% and classifier has an accuracy of 86.6%.

Khurram Shahzad and Imran Siddique [4], paper titled “Efficient Brain Tumor
Detection Using Image Processing Techniques “focused on an easy, fully automatic and
efficient algorithm for extraction of brain tumor has been introduced. Morphological operation
like erosion and dilation along with morphological gradient and threshold are used.
Morphological gradient is used for calculating threshold. Threshold is used to binarize the image
which results an image having tumor and some noise with it. Erosion is used for thinning the
image as it shrinks the image and helps to reduce noise or unwanted small objects. Dilation is
being used after erosion so that to get removed tumor portion back which was being removed by
erosion.

Department of CSE, JSSATEB 2024-2025 Page 4

Brain Tumor Detection Using MobileNet Architecture

Badran et al. [5] adopted the canny edge detection model accumulated with Adaptive
thresholding to extract the ROI. The dataset contained 102 images. Images were first
preprocessed, then for two sets of a neural network, for the first set canny edge detection was
applied, and for the second set, adaptive thresholding was applied.

Pei et al. [6] proposed a technique which utilizes tumor growth patterns as novel
features to improve texture based tumor segmentation in longitudinal MRI. Label maps are
being used to obtain tumor growth modeling and predict cell density after extracting textures
(e.g., fractal, and mm) and intensity features.

Concentrating on Region based Fuzzy Clustering and deformable model, Rajendran

et al. [7] accomplished 95.3% and 82.1% of ASM and Jaccard Index based on Enhanced
Probabilistic Fuzzy C-Means model with some morphological operations.

Department of CSE, JSSATEB 2024-2025 Page 5

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 3
SYSTEM REQUIREMENT SPECIFICATION

3.1 HARDWARE REQUIREMENTS

Hardware requirements for the implementation and testing of this project are mentioned
as below:
• Processors: Intel Atom®processororIntel®Core™i3processor
• Diskspace: 1GB or more
• Ram:4Gb or more
• Operating systems: Windows10/11/

3.2 SOFTWARE REQUIREMENTS

Software requirements for the implementation and testing of this project are mentioned as below.
• Visual Studio Code: For this project, Visual Studio Code is a lightweight but powerful source
code editor which runs on your desktop and is available for Windows, macOS and Linux. It
comes with built-in support for JavaScript, Typescript and Node.js and has a rich ecosystem of
extensions for other languages and runtimes (such as C++, C#, Java, Python, PHP, Go, .NET).

• Python: Python is a general-purpose interpreted, interactive, object-oriented, and high-level

programming language. Python is designed to be highly readable. It supports functional and
structured programming methods as well as OOP. It can be used as a scripting language or can
be compiled to byte-code for building large applications.

3.3 FRAMEWORKS AND LIBRARIES

• NumPy: NumPy is a fundamental library for numerical computations in Python. It provides

support for handling multi-dimensional arrays and various mathematical operations, which are
essential for data preprocessing and manipulation in the brain tumor detection system.

Department of CSE, JSSATEB 2024-2025 Page 6

Brain Tumor Detection Using MobileNet Architecture

• Matplotlib: Matplotlib is a plotting library that provides functionalities for creating

visualizations and plots. It can be used in the brain tumor detection system to generate
visualizations of tumor heatmaps or bounding boxes, enabling better interpretation, and
understanding of the results.

• TensorFlow: The brain tumor detection use TensorFlow for model training and inference with
the MobileNet architecture. These frameworks offer a wide range of tools and utilities for
building, training, and deploying deep learning models efficiently.

• Scikit-learn: Scikit-learn is a comprehensive machine learning library that provides a wide

range of algorithms and tools for data preprocessing, feature extraction, and evaluation. It can
be utilized in the brain tumor detection system for tasks such as dataset splitting, performance
evaluation, and metrics calculation.

• Keras: Keras is a high-level deep learning library that provides a user-friendly interface for
building and training neural networks. It can be used in conjunction with TensorFlow or other
backend frameworks to facilitate the implementation of the MobileNet architecture and
streamline the training process.

Department of CSE, JSSATEB 2024-2025 Page 7

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 4

SYSTEM DESIGN

4.1 MOBILENET ARCHITECTURE

Fig 1. MobileNet Architecture

Transfer learning is a machine learning method that uses previously trained model. So, we have
used a pre-trained CNN architecture called mobile net which is open sourced by google. Mobile
Net uses depth wise separable convolutions to build a light weight deep neural networks and
provides an efficient model. This architecture allows the mobile net to detect and classify brain
tumor efficiently while maintaining a high-level accuracy. Here is an explanation of the key
components and features of MobileNetV2:

• Depth wise Separable Convolution: MobileNetV2 extensively employs depth wise separable
convolutions. In a traditional convolutional layer, the convolution operation is applied to all
input channels, resulting in a high computational cost. In depth wise separable convolutions, the
operation is split into two separate steps: depth wise convolution and point wise convolution.
Depth wise convolution applies a single filter to each input channel independently, reducing the
number of parameters and computation. Point wise convolution then applies a 1x1 convolution
to combine the outputs of the depth wise convolution. This factorization significantly reduces
model size and computational complexity.

Department of CSE, JSSATEB 2024-2025 Page 8

Brain Tumor Detection Using MobileNet Architecture

• Inverted Residuals: MobileNetV2 introduces a new building block called the inverted residual
block, which aims to capture both low-level and high-level features efficiently. The inverted
residual block consists of three main components: a 1x1 pointwise convolution to expand the
number of channels, a depth wise separable convolution to capture spatial information, and
another 1x1 pointwise convolution to reduce the number of channels back to the desired output
size.

. Linear Bottlenecks: To further reduce computational complexity, MobileNetV2 adopts linear

bottlenecks in its inverted residual blocks. The linear bottleneck consists of a non-linear
activation function followed by a 1x1 pointwise convolution. This design choice helps preserve
information and prevent the bottleneck layer from becoming a computational bottleneck in terms
of memory and runtime.

• Inverted Residual with Expansion Layer: MobileNetV2 introduces an expansion layer in the
inverted residual block to increase the capacity of the network without significantly increasing
the number of parameters. The expansion layer applies a 1x1 point wise convolution to expand
the number of channels before the subsequent depth wise separable convolution. This expansion
layer allows the network to learn more complex representations while maintaining efficiency.

• Resolution Multiplier: MobileNetV2 also incorporates a resolution multiplier parameter,

which scales the input resolution of the model. Reducing the input resolution decreases both
memory and computational requirements, making the model more efficient. However, this
reduction may also impact the model's ability to capture fine-grained details and small objects.
By combining these design elements, MobileNetV2 achieves a good balance between model
size, computational efficiency, and accuracy. It is particularly well-suited for deployment on
resource-constrained devices or applications that require real-time image classification.

Department of CSE, JSSATEB 2024-2025 Page 9

Brain Tumor Detection Using MobileNet Architecture

4.2 METHODOLOGY
The methodology outlined below provides a general framework for conducting the study on
brain tumor detection using the MobileNet architecture.

Fig 2. System Architecture and dataflow diagram

Department of CSE, JSSATEB 2024-2025 Page 10

Brain Tumor Detection Using MobileNet Architecture

Data Collection: The dataset used for training and validation was collected from the Kaggle. It
contains brain MRI Image’s in which some of them containing tumor (with tumor) and some are
normal (without tumor). In which tumor images are segregated in the folder name “Brain tumor”
with 1500 images and normal images are kept in “healthy” with 1500 images . In this study thus
we have used 3000 Images out of these images 1500 images are with tumor and 1500 images
are without tumors. For testing we have used 256 Images. Fig 3(a) and Fig 3(b) illustrate the
sample brain MRI images with and without brain tumors respectively .

Fig 3(a). Dataset From Kaggle

Fig 3(b). Dataset From Kaggle

• Data Preprocessing: Preprocess the MRI images to prepare them for training the MobileNet
model. This includes resizing the images to a consistent resolution, normalizing pixel values,
and converting them to a suitable format compatible with the deep learning framework.

• Model Architecture: Implement the MobileNet architecture using a deep learning framework
such as TensorFlow. Customized the output layer(s) of the MobileNet model to correspond to
the brain tumor detection task.

Department of CSE, JSSATEB 2024-2025 Page 11

Brain Tumor Detection Using MobileNet Architecture

• Training Setup: Split the preprocessed dataset into training and validation sets using a
predefined split ratio. Set up the training parameters, such as batch size "32" number of epochs
"10", and optimizer "Adam", to train the MobileNet model.

• Model Training: Trained the MobileNet model using the training set. During training, feed
batches of preprocessed images through the model, compute the loss using an appropriate loss
function "binary cross-entropy" and backpropagate the gradients to update the model's weights.

• Model Evaluation: Evaluate the trained MobileNet model on the validation set. To Compute
evaluation metrics such as accuracy to assess the model's performance in detecting brain tumors.

Department of CSE, JSSATEB 2024-2025 Page 12

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 5

SYSTEM IMPLEMENTATION

5.1 MODEL TRAINING SETUP:

The model training setup consists of several key components, including dataset split,
preprocessing, model initialization, fine-tuning, loss function, training process, hypermeter
tunning, model evaluation.

5.2 DATASET SPLIT:

The brain tumor dataset obtained from Kaggle dataset is typically divided into three
subsets: training set, validation set, the training set is used to train the model, the validation set
is used for hyper parameter tuning and model evaluation during training, and the test set is used
to evaluate the final model's performance. The experiments are repeated for different split ratio
(train set: validation set: 2700:300) i.e. 9: 1, (train set: validation set:2400:600) i.e. 8: 2 and (
train set: validation set:: 2100 : 900 ) i.e. 7: 3 . The test set has 256 images.

5.3 PRE-PROCESSING:
Preprocessing steps are applied to the MRI images before feeding them into the
model. These steps may include resizing the images to a consistent resolution, normalization,
and converting the images into the appropriate format compatible with the chosen deep learning
framework. The image size is set to 224X224.

5.4 MODEL INITIALIZATION:

The MobileNet architecture, pre-trained on a large-scale dataset (such as Image Net),
can be used as the starting point for the brain tumor detection task. The pre-trained weights of
MobileNet are loaded into the model, which helps in capturing relevant image features. This
process is known as transfer learning.

Department of CSE, JSSATEB 2024-2025 Page 13

Brain Tumor Detection Using MobileNet Architecture

5.5 FINE-TUNING:
In addition to transfer learning, fine-tuning can be performed by freezing some
initial layers of the MobileNet model while allowing the subsequent layers to be fine-tuned using
the brain tumor dataset. This approach enables the model to adapt to the specific characteristics
of brain tumor images while leveraging the learned features from the pre-trained model.

5.6 LOSS FUNCTION AND OPTIMIZATION:

A suitable loss function, which is binary cross-entropy is chosen based on the nature
of the brain tumor detection task binary class. The model is optimized using an optimization
algorithm Adam, with appropriate learning rate and regularization techniques (e.g., weight
decay, dropout) to prevent over fitting.

5.7 TRAINING PROCESS:

The model is trained using the training set, with batches of images and their
corresponding labels. The training process involves feeding the images forward through the
model, calculating the loss, and back propagating the gradients to update the model's weights.
The process is repeated for multiple epochs until convergence or a predefined stopping criterion
is met

5.8 HYPERPARAMETER TUNING:

During the training process, hyper parameters such as learning rate, batch size,
number of layers to freeze, regularization strength, and optimizer parameters can significantly
impact the model's performance. These hyper parameters are tuned using the validation set to
find the optimal combination that maximizes the model's accuracy or other desired evaluation
metrics.

Department of CSE, JSSATEB 2024-2025 Page 14

Brain Tumor Detection Using MobileNet Architecture

5.9 MODEL EVALUATION:

Once the model training is completed, the trained model is evaluated on the test set,
which contains previously unseen brain tumor images. Performance metrics such as accuracy,
Precision , Recall and F1 Score are calculated to assess the model's effectiveness in detecting
brain tumors.

5.10 SIGMOID ACTIVATION FUNCTION

The sigmoid function is used as the activation function in the output layer of the
model. The MobileNetV2 architecture, like many other architectures for binary classification
tasks, typically employs a sigmoid activation function in the final layer. The purpose of the
sigmoid function is to produce a probability-like output that represents the likelihood of a brain
tumor being present in a given input image. The sigmoid function takes the weighted sum of the
inputs and applies a non-linear transformation, squashing the output between 0 and 1.
In the MobileNetV2 architecture for brain tumor detection, the output layer consists of a single
neuron with a sigmoid activation function. During inference, the output of this neuron represents
the probability of the input image containing a brain tumor. Values close to 0 indicate a low
probability of tumor presence, while values close to 1 indicate a high probability.
By utilizing the sigmoid activation function, the MobileNetV2 model can provide a probabilistic
output,which is particularly useful in binary classification tasks such as distinguishing between
tumor and non-tumor regions in brain MRI scans.

5.11 IMPLEMENTATION ALGORITHM

An algorithm is a procedure or formula for solving a problem, based on conducting a
sequence of specified actions. Algorithms are usually used throughout all areas of IT amount of
space and time and in a well-defined formal language for calculating a function .As a effective
method algorithm can be expressed with in a finite. Fig Algorithm is about how the system
works on whole with inputting image and preprocessing then training using MobileNet model
and output shows an healthy brain or tumor detected.

Department of CSE, JSSATEB 2024-2025 Page 15

Brain Tumor Detection Using MobileNet Architecture

Step 1: Start
Step 2: Prepare Dataset (Healthy / Unhealthy)
Step 3: Preprocessing
Target Size (224,224)
Step 4: Train MobileNet Model
Step 5: Input image from PC
Step 6: Preprocessing
Target Size (224,224)
Step 7: Test model
Step 8: If result <= 0.5:
Result = “the person has no brain tumor”
Else:
Result = “the person has brain tumor”

Fig 4. Algorithm for implementation

Department of CSE, JSSATEB 2024-2025 Page 16

Brain Tumor Detection Using MobileNet Architecture

CHAPTER6
RESULTS AND DISCUSSIONS
The result forms the most important part of the project. This section provides the outcomes of
the project.
6.1 TRAINING & VALIDATION ACCURACY AND LOSS
Table 2 shows the comparison on training and validation accuracy from 0.1 to 0.3 based on
Validation split.

Table 1: Training & loss comparison

Training accuracy refers to the accuracy of the model during the training phase. It measures the
percentage of correctly classified instances from the training dataset. The training accuracy
indicates how well the model is fitting the training data and learning the patterns associated with
brain tumor detection.
Validation accuracy is the accuracy of the model on a separate validation dataset. This dataset is
not used for training but is used to evaluate the model's performance during training. The
validation accuracy provides an estimate of how well the model can generalize to unseen data..
Below fig 5 & fig 6 shows training and validation graph

Fig 5. Model Accuracy Fig 6. Model Loss

Department of CSE, JSSATEB 2024-2025 Page 17
Brain Tumor Detection Using MobileNet Architecture

6.2 PERFORMANCE METRICS

The confusion matrix is a valuable tool for evaluating the performance of a classification
model, such as the brain tumor detection system using MobileNet. It provides a detailed
breakdown of the model's predictions and can help assess its accuracy, precision, recall, and
other performance metrics.
Fig 7. shows the confusion matrix which is having accuracy of 95.25 %

Fig 7. Confusion matrix

• Accuracy:
The accuracy of the brain tumor detection system was found to be 94.86%. This indicates
the overall percentage of correctly classified tumor and non-tumor regions. A high accuracy
suggests that the model is effectively identifying the presence or absence of tumors.
Accuracy is calculated by:
Accuracy = (True positive + True Negative) / Total no of images
Accuracy = (97+144) / 253
Accuracy = 241/ 253 = 95.25 %.

• Precision:
The precision of the system was calculated as 95.25%. This metric measures the
Proportion of correctly predicted tumor regions out of all predicted tumor regions. A high

Department of CSE, JSSATEB 2024-2025 Page 18

Brain Tumor Detection Using MobileNet Architecture

precision indicates a low number of false positives, meaning the system is correctly identifying
tumor regions and minimizing the risk of misdiagnosis.

• Recall:
The recall, also known as sensitivity or true positive rate, was determined to be 94.86%.
This metric measures the proportion of correctly predicted tumor regions out of all actual tumor
regions. A high recall indicates that the system is effectively detecting true positive tumor
regions and minimizing false negatives.

• F1 Score:
The F1 score, which is the harmonic mean of precision and recall, was calculated as 94%.
This metric provides a balanced measure of the model's accuracy, taking into account both false
positives and false negatives. A high F1 score indicates a well-performing system with a good
balance between precision and recall.
Overall, the results from the confusion matrix analysis support the effectiveness of
the brain tumor detection system using MobileNet. The system demonstrates promising
performance metrics, but further research and improvements may be necessary to enhance its
robustness and generalizability to real-world scenarios.
Table 3 shows the comparison of Performance metrics for various split ratios ranging from 0.1
to 0.3.Split ratio is nothing but number of images used for validation to total images used for
both training and validation .

Table 2: Performance Metrics comparison

Department of CSE, JSSATEB 2024-2025 Page 19

Brain Tumor Detection Using MobileNet Architecture

6.3 SNAPSHOTS

Department of CSE, JSSATEB 2024-2025 Page 20

Brain Tumor Detection Using MobileNet Architecture

Department of CSE, JSSATEB 2024-2025 Page 21

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 7
CONCLUSION
In this project work, we presented a comprehensive study on brain tumor detection using the
Mobile Net model. It is aimed to develop an efficient and accurate solution for detecting brain
tumors. Through dataset acquisition, preprocessing, model training, and evaluation, the
effectiveness of the Mobile Net model is accurately demonstrated by identifying brain tumors
from medical images( MRI).
Expanding the dataset with a larger and more diverse sample can enhance the generalizability of
the Mobile Net model. Additionally, exploring transfer learning techniques, incorporating
clinical data, and investigating multi-modal fusion could further enhance the model's
performance and contribute to personalized medicine in brain tumor diagnosis and treatment.

Department of CSE, JSSATEB 2024-2025 Page 22

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 8

FUTURE SCOPE
The project aimed to develop an efficient and accurate solution for detecting brain tumors.
Through dataset acquisition, preprocessing, model training, and evaluation, we demonstrated the
effectiveness of the Mobile Net model in accurately identifying brain tumors from medical
images.
Integrating the brain tumor detection system with decision support systems can assist healthcare
professionals by providing additional information and recommendations. For example, the
system can provide information on tumor characteristics, growth patterns, and treatment options,
aiding in the decision-making process.

Department of CSE, JSSATEB 2024-2025 Page 23

Brain Tumor Detection Using MobileNet Architecture

CHAPTER 9
REFERENCES

[1] Dr. Chinta Someswara rao “Brain Tumor Detection Model from MR Images using
Convolutional Neural Network” IEE May [June 2020]

[2] Masoumeh Siar “Brain Tumor Detection Using Deep Neural Network and Machine Learning
Algorithm” International Conference on Computer and Knowledge Engineering [October 2019]

[3] B. Devkota, Abeer Alsa doon, P.W.C. Prasad, A. K. Singh, A. Elchouemi, “Image
Segmentation for Early-Stage Brain Tumor Detection using Mathematical Morphological
Reconstruction,” 6th International Conference on Smart Computing and Communications,
ICSCC 2019, 7-8 December 2019, Kurukshetra, India.

[4] Khurram Shahzad and Imran Siddique “Efficient Brain Tumor Detection Using Image
Processing Techniques “International Journal of Scientific & Engineering Research [December
2019]

[5] Ehab F. Badran, Esraa Galal Mahmoud, Nadder Hamdy, “An Algorithm for Detecting Brain
Tumors in MRI Images”, 7th International Conference on Cloud Computing, Data Science &
Engineering - Confluence, 2020.

[6]Pei L, Reza SMS, Li W, Davit zikrs C, Iftekhar Uddin KM. “Improved brain tumor
segmentation by utilizing tumor growth model in longitudinal brain MRI”. Proc SPIE Int Soc
Opt Eng. 2021.

[7] A. Rajendran, R. Dhanasekaran, “Fuzzy Clustering and Deformable Model for Tumor
Segmentation on MRI Brain Image: A Combined Approach,” International Conference on
Communication Technology and System Design 2020.

[8]Aryan Sagar Methyl “Brain Tumor Detection using Deep Learning and Image Processing
“IEEE [June 2021]

[9] Dataset https://www.kaggle.com/navoneel/brain-mri-images-for-brain-tumor-detection

Department of CSE, JSSATEB 2024-2025 Page 24