An Empirical Study of Brain Tumor Classification Using MRI Images

Brain tumor (BT) arises due to uncontrollable and fast development of cells. Unless addressed at a primary stage, it may cause death. The information about the human soft tissues for BT classification is gathered by Magnetic resonance imaging (MRI). BT classification is a complicated chore because different parts of the same tumor can exhibit diverse characteristics (e.g. texture, density), making it difficult to classify accurately. In order to overcome the issues, this survey provides an analysis of existing BT classification approaches. This study examines 50 research papers focused on several approaches utilized for BT classification. Initially, this survey demonstrates the MRI image-based BT classification process. A detailed discussion about the five key techniques involved in classification methods like Transfer Learning (TL)-based methods, Deep learning (DL)-based models, Machine Learning (ML)-based techniques, Algorithmic-based approaches, and Hybrid methods are provided. Additionally, it describes the BT classification method using MRI images and research gaps. Finally, the evaluation is provided on the basis of performance evaluation, research methodologies, publication year, and achievement of the research methods. The analysis shows that most of the papers employed DL methods for BT classification. Likewise, the frequently utilized dataset is the BT dataset in reviewed papers. Similarly, accuracy is the most widely used evaluation measure.