Skip to main content
. 2020 Aug 4;19(2):267–284. doi: 10.1007/s12021-020-09483-7

Fig. 6.

Fig. 6

Quantitative evaluations of different registration methods. Heat map of 140 brain regions or nuclei. The applied metric is Dice scores. The closer a color is to red, the lower the registration accuracy is, and the closer a color is to blue, the higher the registration accuracy is. Quantitative evaluations of Affine (0.493 ± 0.200), LD demons (0.637 ± 0.136), SyN (0.756 ± 0.106), QuickSilver (0.728 ± 0.103), VNet (0.740 ± 0.122) and DeepMapi (0.825 ± 0.081) are presented. Additionally, an asterisk (*) is used to mark the regions with the highest DeepMapi values