PDF version of this entire document

Brain Dataset with Ground Truth

Our initial dataset consisted of 74#74 transaxial mid-brain 2D slices, extracted at equivalent levels from a set of T1-weighted 3D MR scans of normal subjects. The ground-truth data for this set consisted of dense (pixel by pixel) binary tissue labels, the tissue classes being gray and white matter, the caudate nucleus, and CSF within the lateral ventricles. These labels were further divided into left and right. An example image and it's labelling is shown in Figure .

Figure: An example affinely-aligned brain image and its accompanying anatomical labels, both overlaid and expanded, for gray matter, white matter, the lateral ventricles, and the caudate nucleus. Labels are also divided into left and right.

75#75

The training set was non-rigidly registered using the Minimum Description Length (MDL) algorithm [#!IPMI_2005_ISBE!#]. This registration was used as the starting point for the evaluation.