Table 5.

Results on the testing set from dataset1. Results are presented as mean ± standard error of the mean across the dataset. Data augmentation does not apply to FreeSurfer and FastSurfer. Volume Error Rate denoted as VER, absolute VER denoted as AVER. Best performances are denoted in boldface.

Method	Data Augmentation	Dice	Recall	Precision	VER	AVER	Pearson r
1-step	no	0.73 ± 0.01	0.72 ± 0.01	0.75 ± 0.01	0.07 ± 0.03	0.18 ± 0.02	0.84
	yes	0.71 ± 0.01	0.68 ± 0.02	0.76 ± 0.01	0.17 ± 0.04	0.25 ± 0.04	0.77
2-step	no	0.72 ± 0.01	0.70 ± 0.01	0.76 ± 0.01	0.12 ± 0.03	0.19 ± 0.02	0.86
	yes	0.69 ± 0.01	0.65 ± 0.01	0.76 ± 0.01	0.20 ± 0.04	0.26 ± 0.04	0.78
FreeSurfer		0.33 ± 0.01	0.42 ± 0.01	0.29 ± 0.01	−0.29 ± 0.03	0.32 ± 0.02	0.65
FastSurfer		0.35 ± 0.01	0.40 ± 0.01	0.32 ± 0.01	−0.18 ± 0.03	0.25 ± 0.02	0.59