A Critical Analysis of Brain Tumor MRI Segmentation and Classification Utilizing Machine Learning and Deep Learning Methods

Abstract

Brain Tumor (BT) result from uncontrolled cell growth and can be fatal if not treated. Classification and segmentation of data remain difficult despite many large-scale initiatives and encouraging results. Variations in the location, shape, and size of tumors make diagnosis difficult for doctors. This report provides a comprehensive literature analysis on magnetic resonance imaging (MRI) to aid researchers in detecting brain tumors (BT). The subject matter includes topics such as the anatomy of the Brain Tumor (BT), publicly available datasets, methods to improve the quality of images, dividing the tumor into distinct parts, extracting important characteristics, categorizing the tumor, using advanced Machine Learning (ML) techniques like Deep Learning (DL), transferring knowledge from one task to another, and employing fuzzy sets for analysis. The review provides an extensive overview of ML and DL methods for BT classification. With its capability to analyse vast amounts of data, DL has shown outstanding performance across various fields, particularly in biomedicine. This assessment offers comprehensive information about both 2-dimensional (2D) and 3-dimensional (3D) datasets and the methodologies used. The use and testing of MRI scans have been utilized to identify BT, resulting in positive outcomes. The goal of this study is to conduct a thorough and critical review of existing research on BT detection and classification using MRI.