Segmentation of Magnetic Resonance Images Using Fuzzy Clustering

Magnetic Resonance Imaging (MRI) is unique among imaging modalities in that image contrast can be varied over a wide range by manipulation of the acquisition parameters. The signal intensity depends indirectly on particular physical and chemical characteristics of the tissues being imaged. These tissue properties influence the behavior of the nuclei undergoing resonance, and their behavior is what directly affects the MRI signal. The parameters of interest that describe this behavior are the relaxation times Ti and T2, the spin density (for hydrogen, N(H)), and the microscopic (diffusion) and macroscopic (flow, motion) motional states of the nuclei. These parameters exist, in turn, because of certain properties of the tissues: water content, fat content, macromolecules, paramagnetic ions and flow being among the significant variables.