AN APPLICATION OF INTEGRATED CLUSTERING TO MRI SEGMENTATION

Magnetic Resonance Imaging (MRI) plays a relevant role in the design o f systems for computer assisted diagnosis. MR-images are multi-dimensi onal in nature; radiologists have to combine several perceptual inform ation coming from more than one image (usually less-than-or-equal-to 4 ). This to perform the tissue classification needed for diagnosis. Aut omatic clustering methods help to discriminate relevant features and t o perform preliminary segmentation of an image; it can guide the manua l classification of body-tissues. Here four clustering techniques and their integration, in an information fusion clustering procedure, are described. The accuracy of the methodology considered is evaluated on real image data. The evaluation is based on the comparison of the resu lts obtained by automatic procedures with the tissue-classification do ne by radiologists.