. Author manuscript; available in PMC: 2022 Jun 1.

Published in final edited form as: Artif Intell Med. 2021 Apr 23;116:102078. doi: 10.1016/j.artmed.2021.102078

Table 4.

Test accuracy on CP-younger fetus and CP-older fetus datasets for models learned with R.I. and with T.L. via fine-tuning a model pre-trained on CP-newborn.

				DSC	HD95 (mm)

Figure 1	CP- younger fetus	n train=16	R.I.	0.90 ± 0.03	0.80 ± 0.02
		n train=16	T.L.	0.90 ± 0.03	0.80 ± 0.01

		n train=5	R.I.	0.88 ± 0.03	0.86 ± 0.12
		n train=5	T.L.	0.89 ± 0.03	0.83 ± 0.06

	CP- older fetus	n train=10	R.I.	0.82 ± 0.05	1.02 ± 0.19
		n train=10	T.L.	0.82 ± 0.05	0.96 ± 0.19

		n train=5	R.I.	0.79 ± 0.05	1.20 ± 0.26
		n train=5	T.L.	0.82 ± 0.04^*	0.97 ± 0.20^*

UNet++	CP- younger fetus	n train=16	R.I.	0.90 ± 0.04	0.92 ± 0.06
		n train=16	T.L.	0.91 ± 0.05	0.87 ± 0.04

		n train=5	R.I.	0.90 ± 0.04	0.99 ± 0.10
		n train=5	T.L.	0.91 ± 0.03	0.84 ± 0.06^*

	CP- older fetus	n train=10	R.I.	0.83 ± 0.03	1.15 ± 0.16
		n train=10	T.L.	0.84 ± 0.04	1.10 ± 0.17

		n train=5	R.I.	0.80 ± 0.05	1.41 ± 0.28
		n train=5	T.L.	0.84 ± 0.02^*	1.05 ± 0.18^*

Asterisk (*) denote statistical significance due to T.L.