Analyzing and classifying MRI images using robust mathematical modeling

Medical imaging is an exponentially growing field, which consists of a set of tools and techniques used to extract useful information from medical images. Magnetic Resonance Imaging (MRI) is one of the most popular techniques among image modalities. This paper develops a linear model for classifying MRI images into the tumor and non-tumor categories. The proposed algorithm supports automatic extraction of features from brain MRI images, and focuses on extracting grey matter and white matter, so that the unhealthy MRI images can be isolated from the healthy MRI images. This technique takes advantage of preprocessing strategies and various filters for viable extraction and for classifying the brain MRI images. The samples of MRI images are taken from the BRAINIX and Neuroimaging data sources. The results are validated by applying the mathematical equations on 46 patients categorizing into 24 subjects as healthy and the remaining 22 as unhealthy. The novelty lies in formulating a general equation for both groups, which can be further used as a tool in computer-assisted medical systems for classifying patients.