Table 1.

Comparison of different segmentation methods.

Method	DSC (%)	IOU (%)	Recall (%)	Precision (%)	F1 Score (%)
FCN	59.31 ± 9.30	44.60 ± 9.32	66.07 ± 7.93	57.69 ± 9.69	59.31 ± 9.30
U-Net	62.12 ± 8.91	47.84 ± 9.42	68.89 ± 8.44	60.44 ± 8.64	62.12 ± 8.91
U-Net++	61.60 ± 9.26	47.48 ± 9.73	67.43 ± 6.60	60.37 ± 11.26	61.60 ± 9.26
Res U-Net	63.93 ± 8.82	50.32 ± 9.31	62.71 ± 9.20	71.37 ± 7.66	63.93 ± 8.82
3D U-Net	61.45 ± 8.64	49.41 ± 10.18	58.64 ± 12.91	65.58 ± 8.73	61.45 ± 8.64
SW-3DUNet	51.18 ± 9.87	36.26 ± 9.08	60.68 ± 12.10	47.62 ± 10.17	51.18 ± 9.87
SANet	64.33 ± 8.68	50.59 ± 9.31	63.09 ± 8.23	70.90 ± 8.40	64.33 ± 8.68
DBR-Net	51.14 ± 7.57	36.71 ± 6.77	48.88 ± 7.69	59.99 ± 6.03	51.14 ± 7.57
Proposed method	66.83 ± 9.06	56.33 ± 7.08	61.12 ± 11.9	82.31 ± 6.39	66.83 ± 9.06

¹The variable was expressed as the mean ± standard deviation (SD). The bold value represents the optimal result for the column of indicators.