Systematic Report on Exploration of Machine
Learning Methods for Brain Tumor Prediction
1
Nadenlla RajamohanReddy
2*
G. Muneeswari
School of Computer Science and Engineering School of Computer Science Engineering
VIT-AP University VIT-AP University
Amaravati, Guntur District, A.P Amaravati, Guntur District, A.P
rajamohan.22phd7007@vitap.ac.in muneeswari.g@vitap.ac.in
Abstract— Recently, there has been a lot of interest among
researchers in machine learning methods that use computational
modules to build information representations. The most fragile
and important organ in the human body is the brain. The
imagination, hearing, listening, perception, and other senses are
under the control of the central nervous system. According to
statistics from the World Health Organization, brain tumors will
account for roughly 10 million deaths globally in 2020. This article
offers a thorough examination of current strategies that deliver
the highest results in their respective industries. The proposed
CNN model using Pooling algorithm such as Average Pooling, Max
Pooling, and Global Average Pooling. Different classifiers can be
used to extract useful features from the MRI images.
Keywords— Brain Tumor, Machine Learning, Feature
extraction, Magnetic Resonance Images, Convolutional Neural
Network, Pooling Algorithm.
I. INTRODUCTION
Brain tumor occurs as a result of the growth of abnormal
cells. Chemotherapy, radiation therapy, and surgery are
preferred brain tumor treatments. Even though benign tumors
only affect one area, their size and location might make them
potentially fatal [1]. The damage that the tumor causes to the
brain's lobes may change a person's personality. Since the
frontal, temporal, and parietal lobes regulate inhibition,
emotions, mood, judgment, thinking, and behavior, a tumor
in those areas may result in improper social behavior.
The most recent WHO categorization of brain tumors has
shown how widely used it is as a prospective future perspective
in epidemiological research projects and therapeutic
procedures [18]. An initial brain tumor is referred to as a
primary brain tumor. In the United States, primary malignant
tumors ofthe brain are expected to be detected in 25,050 adults
this year (14,280 males and 10,530 women). Less than 1
percent of people will experience this form of tumor in their
lives. 85 to 90 percent of all primary central nervous system
(CNS) malignancies are brain tumors [19]. According to the
International Agency for Research on Cancer (IARC) [17]. In
2020, there will likely be 308,102 primary brain tumor
diagnoses worldwide. A diagnosis of a brain or CNS tumor
will also be given to 5.230 children under the age of 15 in the
US this year [19].
There are secondary brain tumors, often known as brain
metastases, in addition to primary brain tumors. At this point,
the tumor, which had first developed elsewhere in the body, had
progressed to the brain. The most typical tumors to spread to
the brain include those caused by leukaemia, lymphoma,
melanoma, breast, kidney, and lung malignancies. Only
primary adult brain tumors are discussed in this manual. The
10th most common cause of mortality for both men and women
are brain and other nervous system cancer. In the United
States, primary malignant brain and CNS tumors are expected
to claim the lives of 18,990 persons this year (11,020 males
and 7,970 women). In 2020, primary CNS and brain tumor
deaths were expected to total 251,329 people worldwide [19].
It's critical to keep in mind that estimates represent the
survival rates for persons with brain tumors. The estimate is
based on yearly data on the number of Americans who have
this tumor. Every five years, specialists measure the survival
rates. This indicates that the estimate might not take into
account changes in the last five years in the diagnosis or
treatment of brain tumors. Examining characteristic brain
structures, such astumors, and pinpointing their exact position
and orientation are important in order to categorize brain
cancers correctly. The quality of the photos [17] has a big
impact on how well segments work. Poor images, for
instance, can produce undesirable outcomes. The automatic
segmentation of clinical images can support imaging-based
analysis and offer suggestions for diagnostic and therapeutic
operations.
Using further segmentation examples to identify brain
tumors, the available ML techniques are illustrated in this
review research. A HGG and an LGG are both visible in the
topand bottom rows, respectively. Edoema is denoted by the
colors green, necrosis by the color blue, a tumor that is not
improvingaccuracy by the color yellow, and an improving
tumor by the color red. Additionally, this essay seeks to
pinpoint any knowledge gaps in the practices now in use.
The classification of every procedure utilized to complete
the tasks in this study has been excellently done by the
authors. New viewpoints on machine learning-based brain
tumor retrieval techniques are provided in this paper.
Proceedings of the 5th International Conference on Inventive Research in Computing Applications (ICIRCA 2023)
IEEE Xplore Part Number: CFP23N67-ART; ISBN: 979-8-3503-2142-5
979-8-3503-2142-5/23/$31.00 ©2023 IEEE 323
2023 5th International Conference on Inventive Research in Computing Applications (ICIRCA) | 979-8-3503-2142-5/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICIRCA57980.2023.10220764
Authorized licensed use limited to: Charles Darwin University. Downloaded on August 29,2023 at 22:55:20 UTC from IEEE Xplore. Restrictions apply.