A Transfer Deep Generative Adversarial Network Model to Synthetic Brain CT Generation from MR Images

Table 4

SSIM indicators of each model.

Sample	FCM	CycleGAN	WGAN	Dual2DGAN	Dual3D&PatchGAN
1	0.5398	0.6101	0.6705	0.6305	0.8252
2	0.4975	0.7932	0.7732	0.7021	0.7895
3	0.5534	0.7205	0.6250	0.6552	0.8102
4	0.5801	0.7022	0.7532	0.7306	0.8521
5	0.5003	0.6352	0.6609	0.8051	0.8012
6	0.5908	0.6010	0.6920	0.7334	0.7965
7	0.4431	0.6905	0.6679	0.6988	0.7849
8	0.4806	0.5405	0.7001	0.7005	0.8502
9	0.5004	0.6609	0.7203	0.6967	0.8106
Means	0.5207	0.6616	0.6959	0.7059	0.8139
Variance	0.0024	0.0056	0.0022	0.0025	0.0006