Skip to main content
. 2023 Jun 25;40(3):482–491. [Article in Chinese] doi: 10.7507/1001-5515.202302050

表 4. Experimental results of ablation on MVI dataset with different coding networks using domain adaptive denormalization.

在MVI数据集上不同编码网络使用域自适应反标准化的消融实验结果

编码网络 ACC AUC REC PRE F1
ResNet (73.98 ± 1.82)% (74.83 ± 1.05)% (78.45 ± 4.13)% (76.72 ± 0.11)% (77.54 ± 1.97)%
ResNet# (75.00 ± 0.64)% (82.79 ± 2.22)% (88.37 ± 2.90)% (72.25 ± 0.89)% (79.39 ± 0.72)%
ShuffleNet (71.93 ± 2.59)% (77.82 ± 3.27)% (84.67 ± 3.50)% (67.12 ± 4.38)% (70.33 ± 3.19)%
ShuffleNet# (72.65 ± 0.74)% (78.24 ± 1.71)% (83.67 ± 3.20)% (69.14 ± 4.52)% (75.43 ± 1.70)%
Swin Transformer (70.98 ± 1.61)% (71.36 ± 0.44)% (81.36 ± 1.76)% (74.66 ± 1.90)% (77.73 ± 0.33)%
Swin Transformer# (72.44 ± 1.93)% (77.63 ± 1.32)% (87.45 ± 2.27)% (69.37 ± 2.43)% (77.18 ± 2.19)%