Table 7.

Quantitative comparison of different synthesis-based and synthesis-free methods for VS segmentation. ${}^{\#}$ and ${}^{\square }$ denote separated and end-to-end methods for synthesis and segmentation, respectively. ${}^{▵}$ denotes synthesis-free methods for the segmentation task. Bold font highlights the best values obtained by the synthesis-based methods. ${}^{†}$ denotes significant improvement from “T2 only” based on a paired t-test (p-value $<$0.05).

Methods	Dice (%)	ASSD (mm)	HD95 (mm)
${}^{\#}$Pix2Pix [15]	83.75$\pm$15.00	0.78$\pm$0.43	2.60$\pm$2.37
${}^{\#}$PGAN [3]	82.90$\pm$12.31	1.19$\pm$1.09	2.99$\pm$5.52
${}^{\square }$[32]	85.89$\pm$6.50	0.53$\pm$0.14	1.54$\pm$0.59
${}^{\square }$PSCGAN [36]	84.65$\pm$10.07	0.65$\pm$0.39	1.79$\pm$0.87
${}^{\square }$UAGAN [41]	87.30$\pm$8.65	0.50$\pm$0.43	1.65$\pm$0.62
${}^{▵}$T2 only	86.00$\pm$14.79	0.71$\pm$0.53	1.96$\pm$1.42
${}^{▵}$Real ceT1	92.80$\pm$3.83	0.31$\pm$0.13	1.09$\pm$0.24
Ours	89.46$\pm$5.49${}^{†}$	0.42$\pm$0.15${}^{†}$	1.31$\pm$0.59${}^{†}$