This paper presents a fully automated algorithm for segmentation of multipl
e sclerosis (MS) lesions from multispectral magnetic resonance (MR) images.
The method performs intensity-based tissue classification using a stochast
ic model for normal brain images and simultaneously detects MS lesions as o
utliers that are not well explained by the model. It corrects for MR field
inhomogeneities, estimates tissue-specific intensity models from the data i
tself, and incorporates contextual information in the classification using
a Markov random field. The results of the automated method are compared wit
h lesion delineations by human experts, showing a high total lesion load co
rrelation. When the degree of spatial correspondence between segmentations
is taken into account, considerable disagreement is found, both between exp
ert segmentations, and between expert and automatic measurements.