Table 3.

Dice scores

Model	CME	NGE
Ground truth	0.44 [0.27, 0.62]	0.65 [0.59, 0.81]
UNet-Baseline	0.43 [0.36, 0.49]	0.43 [0.34, 0.50]
UNet-Phase	0.50 [0.43, 0.59]${}^{\ast }$	0.44 [0.36, 0.60]
Pix2Pix	0.37 [0.28, 0.46]${}^{\ast }†$	0.29 [0.06, 0.38]
CycleGAN	0.49 [0.41, 0.54]$†$	0.50 [0.44, 0.57]

Data format: median [95% CI]

${}^{\ast }$ and $†$ indicate significant difference ($p<0.01$)