Brain Tumor Detection and Classification in MRI: Technique for Smart Healthcare Adaptation

Brain tumor is termed as an unnatural cell growth and division of brain tissues, which may be cancerous or noncancerous. Computed tomography, magnetic resonance imaging (MRI), single-photon emission computed tomography, magnetic resonance spectroscopy, and positron emission tomography work upon extracting different parameters (location, size, shape) of tumor cells. Technically, classification of brain tumor cells in MRI is judged by a step-by-step approach like preprocessing, segmentation, and classification. A healthcare adaptation system needs to be formulated to overcome all these challenges. It is observed that most of the problems must be resolved at the preprocessing step itself before the image is fed into a classifier for classification of tumor as normal or abnormal tissues. Image segmentation algorithms are preferred to be performed on grayscale images, as the existing difficulties of gray scale images is less than that of color images. MRI processing is greatly benefitted by the enhanced usage of machine learning and deep learning.