I. INTRODUCTION
Brain tumors are neoplasms that represent abnormal cell
growth within the brain and its surrounding structure. Al-
though a relatively rare form of cancer – in comparison to
other cancers – analogous to other cancers, it can be benign
or malignant, and it is poorly diagnosed because neurological
symptoms cannot be detected easily, and specific tests like
biopsies and analysis of brain scans, needs to be conducted
by human experts. The process of diagnoses usually involves
three facets: Consideration of molecular features, analysis of
histological characteristics (microscopic analysis of cells and
tissues from the brain), and anatomical location (lobal regions
of the brain).
Holistically, these tumors are categorized as follows:
1) Meningiomas: Tumors originating from the meninges
surrounding the brain and the spinal cord.
2) Pituitary Adenomas: Tumors that originate in the pi-
tuitary gland at the base of the brain.
3) Gliomas: Tumors that originate from the glial cells
in the central nervous system. Gliomas are the most
common type of brain tumors that are developed, and
are therefore known as “primary brain tumors.”
4) Metastatic Lesions: Tumors that originate from cancer
cells in other parts of the body, and have spread via
metastatis to the brain. Therefore, they are known as
“secondary brain tumors”.
5) Medulloblastomas: Tumors that originate from the
cerebellum, and are most commonly found in children.
6) Acoustic Neuromas / Schwannomas: Tumors that orig-
inate from the Schwann cells insulting nerve fibers.
7) Pinealomas / Pineocytoma: Tumors that originate from
the pineal gland.
According to research published by the American Cancer
Society [1], in 2023 for the state of Ohio in the United States, it
was reported that 24 810 adults (with 14 280 being men, and
10 530 being women) in the United States were diagnosed
with a form of cancerous tumors of the brain and spinal
cord, and a 2020 survey found that globally 308 102 people
were diagnosed. Although a small percentage of the overall
population of the United States and the world, respectively,
these patients are eligible to receive the best care possible.
Usually, treatment paths are patient-specific, depending on
several factors, such as the patient’s current state of health, the
presence of other underlying diseases, and so on. As with other
cancers, treatment is very expensive, and requires the patient
to take significant steps to a complete lifestyle overhaul.
This multifarious approach encompasses, amongst others, a
change of diet, incorporation of exercise, radiation therapy,
immunotherapy, chemotherapy, and molecular therapy.
With the current Artificial Intelligence (AI) revolution in
all fields that the world is experiencing, the medical field is
no exception. Specifically, within the context of the analysis
of medical images for diagnosis and prognosis, Convolutional
Neural Networks (CNNs) are a type of NN architecture
designed for computer vision and image processing assign-
ments. Similar to how a perceptron is modeled after biological
neurons, the CNN is modeled after the cortical preprocessing
regions of the striate cortex (primary visual cortex – V1) and
the prostrate cortex (secondary visual cortex – V2), located
at the occipital lobe at the back of the brain. Analogous to
how computer vision tasks in the pre-Machine Learning (ML)
and pre-Deep Learning (DL) eras used to be concerned with
detecting edges, shapes, and textures, the neurons in this region
are sensitive to discerning these patterns. CNNs have proven
to be invaluable in the medical domain, and many healthcare
facilities are incorporating CNN-based classification systems
together with medical experts for early disease detection, and
recommended treatment plans [2]–[4].
Architecturally, the CNN is described as being composed
of two distinguishable layers:
1) Convolutional Layer: These contain adjustable fil-