Medical Image Segmentation For Brain Tumor
Prognosis
BY
MELISA M TWALA (R245138Y)
AND
CHARMAINE B GUMBO (R248843X)
Chapter 1
1.1 Introduction
Medical image segmentation is important for modern healthcare particularly in the diagnosis and
treatment of brain tumours. In this project, images processing techniques are used to segment
brain scans to test tumor analysis and prognosis. What does this to accurately assess tumor size,
shape, and location? The doctors need to see precisely what the tumor is, shape, location, and the
appropriate space, the exact analysis necessary for treatment planning.
1.2 Background of the Study
Brain tumors are life-threatening diseases that require immediate diagnosis and care. Traditional
models depend on manual image analysis; it takes time and is susceptible to human error. In fact,
digital images and artificial intelligence have allowed for the automation of this process.
Sectioning medical image data into functional regions such as for example, comparing healthy
tissue to tumors is characterized by medical image segmentation. This project aims to provide a
system that simplifies and improves tumor segmentation accuracy.
1.3 Problem Definition
Identifying manual brain tumors from MRI images is slow, subjective and may lead to
misdiagnosis. In particular, doctors struggle with image clarity, tumor boundary identification,
and analysis speed. A reliable, semi-automated system that can accurately segment brain tumours
to obtain more precise prognosis is needed.
1.4 Aim
The aim is to develop a system that segments medical images to provide a basis for diagnosing
brain tumor prognosis
1.5 Objectives
- To improve the quality of images to detect tumors through image preprocessing methods that
help detect more accurate tumor detection.
- To construct a segmentation model that is able to separate tumor from normal brain tissue.
- To analyze segmented tumor regions by measuring their size and shape to support prognosis
support.
- to test and evaluate the accuracy of the segmentation system with standard performance
measures.
-To develop a visual interface that is able to show the segmented tumor areas clearly on the MRI
image and be easily understood.
1.6 Justification
This work is important because it aims to improve early diagnosis and treatment planning for
brain tumors. As medical professionals will be supported by automated tools, patient outcomes
will improve. It also reduces analysis time and can help hospitals with low-income or
radiologists.
1.7 Ethical Considerations
Patient Data Privacy:
Any medical images or information used in this project will be fully anonymous. It protects
privacy and personal dignity by ensuring that patients are identifiable and not identified.
Informed Consent:
If the system uses data of a hospital or medical institution, appropriate consent must be sought.
This means patients or institutions must agree to use their data before it is used in the project.
Non-Maleficence (Do No Harm):
The system is designed to promote rather than replace medical professionals. It is therefore not
appropriate to be used in final medical decisions without the supervision of an expert to avoid
misdiagnoses or possibly harm to patients.
Respect for Healthcare Standards:
The rules and regulations on health data will be followed in all development. This ensures that
the system is compatible with medical practices and legal and professional guidelines.
Accuracy and Fairness:
The model will be carefully checked for bias or inaccurate results. This helps ensure that the
system produces reliable outputs that do not annoy or discriminate on any basis by any group or
to misrepresent patients.
1.8 Summary
This project focuses on establishing a model that can be used to predict for the segmentation of
brain tumors using medical image analysis as a pre-diagnosis vehicle. This will be in preparation
for, segmenting and analyzing MRI images. It will make it possible to detect tumors quicker and
more accurately for doctors and will be the decision-making vehicle for physicians, whilst
maintaining high ethical standards in the handling of medical data.