Image segmentation (IS) and local fractal analyses of MR images

Summary form only given, as follows. Operator independent algorithms were developed for image segmentation of magnetic resonance (MR) images for white and gray matter, ventricles, and pathology. Two local fractal algorithms were also developed and validated using known fractal Takagi surfaces and Mr phantoms. These methods were applied to: (1) 15 chronic MS patients with extensive plaques and periventricular effusions who received multiple MR exams over time; (2) in vivo, and in vitro MR mandible and femur images; and (3) MR studies of 15 schizophrenics (S) and 15 controls (C). MR results correlated with clinical changes in 13/15 MS patients. MR fractal results correlated with changes in trabecular bone density and structure. MR differences (0.05<p) between S and C were observed in the temporal and prefrontal lobes