Skip to main content
. 2022 May 12;12:7868. doi: 10.1038/s41598-022-11852-y

Table 3.

Performance comparison on lung segmentation (mean ± standard deviation).

Network Accuracy Sensitivity Dice AUC BF-Score
U-Net1 0.9923 ± 0.0024 0.9824 ± 0.0078 0.9834 ± 0.0083 0.9818 ± 0.0031 0.9135 ± 0.0851
ET-Net25 0.9868 ± 0.0069 0.9765 ± 0.0104 0.9832 ± 0.0177 0.9911 ± 0.0053 0.9014 ± 0.0940
AEC-Net26 0.9927 ± 0.0019 0.9810 ± 0.0094 0.9843 ± 0.0071 0.9917 ± 0.0038 0.9083 ± 0.0890
CE-Net14 0.9935 ± 0.0019 0.9876 ± 0.0089 0.9852 ± 0.0057 0.9916 ± 0.0038 0.9208 ± 0.0970
Attention U-Net16 0.9922 ± 0.0023 0.9765 ± 0.0112 0.9832 ± 0.0067 0.9908 ± 0.0052 0.9197 ± 0.0848
CPFNet18 0.9895 ± 0.0022 0.9837 ± 0.0083 0.9843 ± 0.0071 0.9907 ± 0.0032 0.9129 ± 0.0466
MultiResUNet7 0.9932 ± 0.0024 0.9903 ± 0.0085 0.9829 ± 0.0071 0.9922 ± 0.0035 0.9183 ± 0.0455
nnUnet43 0.9937 ± 0.0028 0.9907 ± 0.0054 0.9823 ± 0.0078 0.9922 ± 0.0045 0.9164 ± 0.0450
MEA-Net (ours) 0.9942 ± 0.0022 0.9903 ± 0.0103 0.9858 ± 0.0057 0.9923 ± 0.0046 0.9332 ± 0.0362

Significant values are in bold.