Information processing in nuclear magnetic resonance imaging

Application of diffusion tensor MRI to neurological segmentation

“…Applications to volume measurement, three-dimensional (3D) visualization of data, and image feature analysis have been widespread (Condon et al, 1986;Gur et al, 1991;Jernigan et al, 1990;Jolles et al, 1989;Kerteaz et al, 1990;Pfefferbaum and Zipursky, 1991;Raine et al, 1990;Rusinek et al, 1991;Schwarzkopf et al, 1990;Seab et al, 1988;Simon et al, 1986;, and such techniques are now being used for practical applications such as drug trials and following the natural history of disease and response to therapy (Kikinis et al, 1992;Velthuzien et al, 1995;Mitchell et al, 1994Mitchell et al, , 1997Jackson et al, 1993;Vaidyanthan et al, 1995). The most prevalent techniques have used two or more images to enhance the information available from a single image alone and made use of analysis methods based upon cluster classification Cline et al, 1990Cline et al, , 1991Taxt et al, 1992;Brown et al, 1992;Gogahan et al, 1987;Vannier et al, 1985;Hyman et al, 1989;Jungke et al, 1988;Simmons et al, 1994;Bonar et al, 1993;Joliot and Mazoyer, 1993;Liang and MacFall, 1994;Taxt and Lundervold, 1994;Kao et al, 1996;Alfano et al, 1997;Bullmore et al, 1995).…”

Automatic Magnetic Resonance Tissue Characterization for Three‐Dimensional Magnetic Resonance Imaging of the Brain

“…The application of pattern recognition methods to the tissue clusters generated by multispectral image pixel intensities results in a high dimensional feature space that inherently should improve image segmentation [8][9][10][11][12][13] shown recently, FCM techniques may potentially improve the contrast between pathology, such as tumor, hemorrhage, or edema, and normal tissue that have similar MRI intrinsic relaxation [I]. Hence, automatic pattern recognition methods should provide a greater confidence level of consistency in image interpretation of pathology boundaries as compared to supervised approaches [14][15][16].…”