Deep Learning: A Revolutionizing Approach To Brain Tumor Classification Using MRI

Abstract

Brain cancer, caused by tumors formed through the irregular and unchecked proliferation of brain cells, poses significant risks, including permanent brain damage and even death if left unmanaged. The number of individuals impacted by brain tumors (BT) is rising worldwide. Positional accuracy and tumor size play a key role in traditional treatments. Thus, creating an automated and meticulous approach to deliver critical information to healthcare professionals is of utmost importance. The integration of various imaging modalities with machine learning (ML) along with its various subsets like deep learning (DL) has enhanced physicians’ ability to identify tumor types with increased precision and reduced time. This paper aims to provide insights into recently developed systems that utilize these techniques to analyze medical imaging for BT diagnosis. Lastly, this paper discusses the major challenges faced by DL algorithms in BT classification and potential advancements in this field. Lastly the paper discusses the use of YOLOv8 and compares its results with the customized CNN (CCNN) BT classification technique. The accuracies obtained were 98.94% and 96.88% respectively.