Ijarcce 2023 12107
Ijarcce 2023 12107
Ijarcce 2023 12107
Abstract: Brain tumors are one of the most threatening types of tumors in the world. Magnetic Resonance Imaging (MRI),
a popular non-invasive strategy, produces a large and diverse number of tissue contrasts in each imaging modality and
has been widely used by medical specialists to diagnose brain tumors. However, the manual segmentation and analysis
of structural MRI images of brain tumors is an arduous and time-consuming task which, thus far, can only be performed
by a professional neurologist. Our project aims to simplify this process with the help of machine learning algorithms so
as to efficiently detect brain tumors in an MRI using a computer. In this project we use two different machine learning
models to compare their success and loss rates and to also identify which algorithm performs better. The two machine
learning algorithms being used are LeNet-5 and a self-designed model of convolutional neural networks (CNN). The loss
rates of the two models can be compared through our project. In a recent study, researchers developed a model based on
deep learning to analyze data. The model used inputs from a psychological questionnaire to estimate an individual’s
psychological age and so on. We believe this project will have a great significance in the coming future not only in
relevance in the medical industry but in the technological industry as well.
I. INTRODUCTION
In recent times several people are struggling from tumor issues, especially in the brain. Most of the time though patients
are required to wait hours/days in order to receive their test diagnosis which can be crucial for the patient’s life. To this
particular disadvantage we have programmed a pre-trained model that will be able to identify a tumor from an MRI image
when placed in the system. We have used a slightly customized version of LeNet-5, which was proposed by Yann LeCun
in the year 1998. It is a multi-layer convolutional neural network for image classification. The other training model used
in our program is a combination of 4 convolutional layers. Through this program we aim not only to make the process of
tumor detection simpler, but also two compare the loss rates of two different machine learning algorithms. Convolutional
neural networks or covnets are neural networks that share their parameters. They can be represented as a cuboid having
length, width (dimension of the image), and height (RGB channels).
Now imagine taking a small patch of this image and running a small neural network on it, with say, k outputs and represent
them vertically. Now slide that neural network across the whole image, as a result, we will get another image with different
width, height, and depth. Instead of just R, G, and B channels now we have more channels but lesser width and height.
This operation is called Convolution. If the patch size is the same as that of the image it will be a regular neural network.
Because of this small patch, we have fewer weights.
© IJARCCE This work is licensed under a Creative Commons Attribution 4.0 International License 55
IJARCCE ISSN (O) 2278-1021, ISSN (P) 2319-5940
[2] "Deep Learning for Brain Tumor Detection and Diagnosis: A Review" by F. A. Khan, M. A. Imran, and M. S.
Imran, published in the Journal of Medical Systems, vol. 42, no. 9, 2018.
Advantages
This is a review paper that summarizes the current state-of-the-art deep learning methods for brain tumor detection and
diagnosis.
The authors provide a comprehensive overview of the different types of deep learning architectures and their performance
on various datasets.
The paper also highlights the challenges and future directions of research in this field. Disadvantages
The paper is a review, it does not present new research or results.
[3] "Brain Tumor Segmentation using U-Net with Fine-Tuning Pre-Trained Models" by J. Isensee, P. Kickingereder,
W. Wick, M. Bendszus, published in the International Conference on Medical Image Computing and Computer-Assisted
Intervention, 2018.
Advantages
The authors propose the use of U-Net, a convolutional neural network architecture, for brain tumor segmentation, and
fine-tune it using pre-trained models.
The method is evaluated on a dataset of multi-modal MRI scans and the results show that it outperforms other state-of-
the-art methods in terms of accuracy and computational efficiency.
Disadvantages
The method is only evaluated on one dataset, and it would be beneficial to see if the method generalizes to other datasets.
[4] "Brain Tumor Detection using Deep Learning with Python" by A. Rajaraman, published in the Journal of
Medical Imaging and Health Informatics, vol. 8, no. 6, 2018.
Advantages
The authors propose a deep learning-based method for automated brain tumor detection and classification.
The method is evaluated on a dataset of MRI scans and the results show that it is effective in detecting and classifying
brain tumors.
The paper also provides a comparison of the method with other state-of-the-art methods.
© IJARCCE This work is licensed under a Creative Commons Attribution 4.0 International License 56
IJARCCE ISSN (O) 2278-1021, ISSN (P) 2319-5940
This project aims to create a simpler method for brain tumor detection and we can conclude that the project has achieved
its aim. Additional features in this project include a side-by-side comparison of two machine learning models.
A lot of future enhancements can be added to this project. It can be modified to help in identifying x-rays or other form
of scans. It can also be upgraded using open-cv to act as a real-time detector for various other purposes.
© IJARCCE This work is licensed under a Creative Commons Attribution 4.0 International License 57
IJARCCE ISSN (O) 2278-1021, ISSN (P) 2319-5940
[1] "Brain Tumor Detection and Classification Using Deep Convolutional Neural Network (DCNN) and SVM" by R. K.
Singh and R. K. Singh, published in the International Journal of Innovative Research in Computer and
Communication Engineering (IJIRCCE), Vol. 5, Issue 4, April 2017.
[2] "Deep Learning for Brain Tumor Detection and Diagnosis: A Review" by F. A. Khan, M. A. Imran, and M. S. Imran,
published in the Journal of Medical Systems, vol. 42, no. 9, 2018.
[3] "Brain Tumor Segmentation using U-Net with Fine-Tuning Pre-Trained Models" by J. Isensee, P. Kickingereder, W.
Wick, M. Bendszus, published in the International Conference on Medical Image Computing and Computer-Assisted
Intervention, 2018.
[4] "Brain Tumor Detection using Deep Learning with Python" by A. Rajaraman, published in the Journal of Medical
Imaging and Health Informatics, vol. 8, no. 6, 2018.
[5] "Automated Brain Tumor Detection and Classification using Deep Learning" by A. N. Srivastava and K. K. Singh,
published in the International Journal of Computer Science and Mobile Computing, vol. 7, no. 12, 2018.
[6] "Segmenting Brain Tumors from Multi-Modal Magnetic Resonance Images using Deep Learning" by M. K. Kalra,
M. J. R. Shukla, and M. Singh, published in the Journal of Medical Imaging and Health Informatics, vol. 8, no. 6,
2018.
[7] "Deep Learning-Based Brain Tumor Segmentation Using 3D Fully Convolutional Networks" by A. M. P. Menze, B.
Jakab, S. Bauer, J. Kalpathy-Cramer, K. Farahani, J. Kirby, et al., published in the IEEE Transactions on Medical
Imaging, vol. 34, no. 10, pp. 1993-2005, Oct. 2015.
BIOGRAPHY
© IJARCCE This work is licensed under a Creative Commons Attribution 4.0 International License 58
IJARCCE ISSN (O) 2278-1021, ISSN (P) 2319-5940
© IJARCCE This work is licensed under a Creative Commons Attribution 4.0 International License 59