Table 1.

Mean metric values for human vs. human and computer vs. human comparisons.

	Recall	Recall ≥ 10 mm	FP per case	Dice per case	Dice per correspondence	Merge error	Split error
Human vs. Human
MTRA (LiTS)	0.92	0.94	2.6	0.70 ± 0.27	0.72 ± 0.11	11	5
LiTS (MTRA)	0.62	0.85	0.3	0.70 ± 0.27	0.72 ± 0.11	5	12
Computer vs. Human
FCN (MTRA)	0.47	0.75	4.7	0.53 ± 0.37	0.72 ± 0.11	7	13
FCN (LiTS)	0.72	0.86	4.6	0.51 ± 0.37	0.65 ± 0.16	12	14
FCN + RF (LiTS)	0.63	0.77	0.7	0.58 ± 0.36	0.69 ± 0.18	11	10

The parentheses denote the dataset used as a reference for the computation of evaluation metrics.