Table 3.

Numerical comparisons and ablation studies of the WUMinn-MGH cross-center brain T1-to-FA MR translation task

Methods	L1 ↓	SSIM ↑	PSNR ↑	IS ↑
w/o UDA (Isola et al., 2017)	158.5±0.4	0.8264±0.0017	25.49±0.06	9.27±0.14
ASS (He et al., 2021)	132.8±0.3	0.8847±0.0014	31.63±0.04	11.54±0.10
ADDA (Tzeng et al., 2017)	129.2±0.4	0.9131±0.0016	34.85±0.05	13.28±0.12
GAUDA (Cui et al., 2020)	129.0±0.5	0.9174±0.0015	35.24±0.06	13.46±0.13
JGLA (Yarram et al., 2022)	128.6±0.2	0.9205±0.0016	35.22±0.03	13.50±0.12
AC:GST	121.2±0.3	0.9692±0.0014	38.65±0.04	16.42±0.14
AC:GST-${𝒜}_{\mathit{\theta }}$	122.0±0.3	0.9637±0.0014	37.94±0.04	16.05±0.12
AC:GST-C	121.7±0.3	0.9641±0.0014	38.26±0.06	15.97±0.13
BM:GST (Liu et al., 2021d)	122.8±0.3	0.9563±0.0014	37.65±0.04	15.12±0.12
BM:GST-A	123.5±0.4	0.9488±0.0016	37.30±0.06	14.61±0.11
BM:GST-E	122.9±0.2	0.9532±0.0015	37.16±0.04	14.69±0.13
Target Supervised	114.6±0.3	0.9815±0.0013	42.27±0.06	18.83±0.10