Skip to main content
. Author manuscript; available in PMC: 2016 Jun 7.
Published in final edited form as: Hum Brain Mapp. 2012 May 19;34(10):2635–2654. doi: 10.1002/hbm.22092

Figure 6.

Figure 6

Cross-correlation-based label fusion with the subcortical ROI. Kappa and Jaccard statistics are given for voting based on the rank order of correlations between the subcortical ROI in the subject to be segmented and each of the templates in the template library. A majority of voxels are kept for the top “n” hits, where “n” was varied from 3 to 19. Results are shown for both ANIMAL- and ANTs-based MAGeT Brain for the globus pallidus, striatum, and thalamus. Except for the case of the pallidum segmentations using ANTs-based MAGeT Brain, an optimal segmentation is reached when using the top 15 templates based on the cross-correlation metric.