Brain Tumor Segmentation in MRI Images

Human brain is complex thing and identifying a disease at earlier stages is one of the difficult tasks in medical field. The identification; manual segmentation and detection of infected areas in brain MRI images are a tedious and time-consuming task. To reduce this time constraint; neural networks are an ideal in recognizing diseases using scans since there is no need to provide a specific algorithm on how to identify the disease. In earlier years of machine learning the features were designed manually by the domain expert and it required deep understanding and domain specific knowledge to carry out this task. In recent years there are tremendous developments in the field of machine learning especially deep learning. Deep learning made it possible to learn the hierarchical features present in the data. The idea is to use this knowledge in the medical field as well to learn these features spontaneously and automatically segment the MRI images. In this paper, we review brain tumor segmentation in MRI images. In due course, we review the different techniques to segment the MRI images. We discuss as well the challenges involved in these segmentation techniques, and their potential in the future.