Table 4.

Detailed data on the number of learnable parameters and computation time for the proposed method and its three simplified versions.

Model	GM			WM			CSF				Computation Time (10 Epochs)	#Learnableparameters
	DSC	JI	HD	DSC	JI	HD	DSC	JI	HD	MSE
Squeeze U-SegNet	92.05	88.06	3.52	93.37	90.42	2.8	91.65	88.06	2.0	0.006	1 h	768,788
Squeeze U-SegNet with multi-scale input	93.44	89.47	4.81	94.78	91.90	4.1	93.32	90.25	3.0	0.005	1.04 h	860,180
Squeeze U-SegNet with multi global attention	94.32	89.25	5.10	95.12	91.40	4.2	94.24	89.22	3.3	0.004	1.15 h	942,164
Proposed method	95.54	91.09	3.21	96.56	92.05	3.1	94.86	90.29	3.0	0.003	1.12 h	1,030,420