Table 3.

Comparison of DICE scores according to the different acquisition centers.

Center Id	#Images	R1 vs. R2	ECNN	CM-YNet (param.)	CM-YNet (mask)
01	96	079 ± 0.16	0.81 ± 0.16	0.75 ± 0.19	0.81 ± 0.11
02	96	0.79 ± 0.14	0.83 ± 0.15	0.81 ± 0.15	0.83 ± 0.13
04	34	0.75 ± 0.17	0.57 ± 0.23	0.74 ± 0.20	0.83 ± 0.08 *
05	80	0.64 ± 0.17	0.84 ± 0.13	0.81 ± 0.16	0.84 ± 0.10
07	14	0.88 ± 0.15	0.85 ± 0.15	0.73 ± 0.18	0.82 ± 0.14
10	156	0.77 ± 0.16	0.68 ± 0.24	0.79 ± 0.15	0.85 ± 0.10 *
11	140	0.82 ± 0.12	0.87 ± 0.10	0.84 ± 0.10	0.87 ± 0.07
13	60	0.78 ± 0.12	0.86 ± 0.12	0.82 ± 0.13	0.86 ± 0.11
18	20	0.74 ± 0.14	0.51 ± 0.27	0.80 ± 0.13	0.86 ± 0.08 *
20	90	0.78 ± 0.16	0.61 ± 0.27	0.78 ± 0.16	0.83 ± 0.12 *
99	58	0.79 ± 0.13	0.78 ± 0.20	0.83 ± 0.13	0.89 ± 0.09 *
21	98	0.76 ± 0.14	0.80 ± 0.15	0.82 ± 0.17	0.86 ± 0.10
22	283	0.71 ± 0.22	0.59 ± 0.22	0.72 ± 0.19	0.74 ± 0.19 *
Total	1225	0.76 ± 0.17	0.73 ± 0.22	0.78 ± 0.17	0.82 ± 0.14

Images from the centers 21 and 22 are from the independent IMIM dataset in which id 22 corresponds to images acquired with old devices. The highest value for each row is highlighted in bold. * The difference between CM-YNet (mask) and ECNN is statistically significant (p < 0.001).