Table 2 Comparison of WSI classification between CAMCSA and other methods on the TCGA Lung Cancer dataset. CAMCSA/without CSA denotes the experimental results of CAMCSA on TCGA Lung Cancer before introducing the CSA module.
From: Cross-slide augmentation for whole slide image classification based on class activation map
Methods | TCGA Lung Cancer | |
|---|---|---|
ACC | AUC | |
Mean pooling | 0.8677 | 0.9089 |
Max pooling | 0.8254 | 0.8897 |
ABMIL | 0.8831 | 0.9235 |
DSMIL | 0.8998 | 0.9352 |
TransMIL | 0.8895 | 0.9278 |
DTFD-MIL | 0.8979 | 0.9388 |
MHIM-MIL | 0.8998 | 0.9536 |
CAMCSA/without CAM and CSA | 0.8879 | 0.9315 |
CAMCSA/without CAM | 0.8971 | 0.9478 |
CAMCSA/without CSA | 0.9131 | 0.9618 |
CAMCSA | 0.9209 | 0.9724 |
