Cluster analysis techniques are gaining widespread use for segmentatio
n of MRI data, especially for volume measurement and 3-D display purpo
ses. This paper describes four improvements to such techniques: (1) Th
e use of intensity simulations to model cluster plots; (2) Correction
of image nonuniformity; (3) Anisotropic smoothing of data; and (4) Aut
omatic isolation of tissues of interest. Simulation of cluster plots a
llows an informed choice of pulse sequence(s) and acquisition paramete
rs to be made. Correction of image nonuniformity and anisotropic smoot
hing reduce the spread of signal intensity from a single tissue thus p
roducing significantly more compact clusters, whilst the isolation of
tissues of interest prevents overlap of clusters from the tissues of i
nterest with those not under consideration. These techniques may be us
ed to improve the results of cluster analysis or traded off, for examp
le to allow lower signal-to-noise images, shorter repetition time imag
es, or fewer images to be used for segmentation.