Skip to main content
. Author manuscript; available in PMC: 2020 Oct 1.
Published in final edited form as: Neuroimage. 2019 May 24;199:553–569. doi: 10.1016/j.neuroimage.2019.05.033

Table 2:

Mean (and std. dev.) of the coefficient of variation (in %) of structure volumes over all six datasets with all algorithms.

ASEG SAMSEG MALF CNN PSACNN
Mean (Std) Mean (Std.) Mean (Std) Mean (Std.) Mean (Std.)
WM 2.82 (0.71) 3.06 (0.92) 2.30 (0.80) 6.01 (0.98) 1.91 (0.70)*
CT 2.68 (0.67) 1.42 (0.77)* 1.80 (0.82) 4.57 (1.37) 1.85 (0.70)
TH 3.70 (1.32) 1.93 (0.72) 1.37 (0.59) 3.38 (1.35) 0.85 (0.60)*
CA 2.02 (0.60) 2.38 (1.31) 1.80 (1.05) 1.68 (1.33) 1.68 (0.85)
PU 2.62 (1.22) 2.88 (1.78) 1.24 (0.74) 3.08 (1.13) 1.27 (0.86)
PA 4.67 (2.50) 2.10 (2.29) 2.65 (1.41) 5.69 (1.41) 1.17 (0.56)*
HI 2.48 (0.78) 2.23 (3.54) 1.99 (1.60) 6.31 (2.08) 1.00 (0.41)*
AM 6.50 (1.79) 2.44 (2.42) 1.93 (1.28) 3.87 (1.28) 1.54 (0.56)
LV 2.00 (0.82) 2.33 (1.36) 2.66 (2.38) 5.94 (1.87) 2.21 (0.86)

Bold results show the minimum coefficient of variation.

*

indicates significantly lower (p < 0.05, using the paired Wilcoxon signed rank test) coefficient of variation than the next best method.