ABSTRACT
In the bioinformatics, image processing plays a vital role for neurologists in clinical diagnosis. Several image modalities are helpful for the diagnoses and tumor segmentation. Magnetic resonance imaging (MRI) is preferred among all modalities due to its noninvasive nature and better representation of tumors' internal information. MRI also poses other challenges, such as ambient noise, varying ranges of strength and photon emission, making it impossible to track and segment. According to the practitioners, manual segmentation is a complex and time-consuming process. Additionally, it requires significant expertise since the practitioners might not clearly watch the tumor region precisely. Whereas computer-aided techniques are supposed to assist practitioners in the entire process of decision-making. Recent advances in convolutional neural networks have enabled their application to medical image segmentation. This chapter introduces the recent developments in brain tumor segmentation approaches assisted with artificially intelligent techniques, profound learning processes and comparisons on benchmark datasets. Several reported approaches are evaluated on similitude, specificity, flexibility and precision criteria. Finally, persisting challenges are highlighted along with potential solutions.
