Slide-Level Classification on TCGA-BRCA (10-fold cross-validation)
86.6F1 ScoreMean-pooling
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Mean-poolingBackbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 86.6 | 95.2 | |
| Max-poolingBackbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 85.2 | 94.8 | |
| DTFD-MILBackbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 84.9 | 95.1 | |
| TransMILBackbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 84.8 | 94.5 | |
| CLAM-SBBackbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 84.5 | 93.5 | |
| Mean-poolingBackbone=Frozen IN-1K ResNet-502023.03 | 81.8 | 91 | |
| DTFD-MILBackbone=Frozen IN-1K ResNet-502023.03 | 81.6 | 89.5 | |
| TransMILBackbone=Frozen IN-1K ResNet-502023.03 | 80.6 | 88.9 | |
| KNN (Max)Backbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 80.2 | 88.2 | |
| CLAM-SBBackbone=Frozen IN-1K ResNet-502023.03 | 79.7 | 87.9 | |
| KNN (Mean)Backbone=Fine-tuned ResNet-50 (Proposed)2023.03 | 78.4 | 90.7 | |
| RNN-MILBackbone=Frozen IN-1K ResNet-502023.03 | 77.6 | 87.1 | |
| DS-MILBackbone=Frozen IN-1K ResNet-502023.03 | 77.5 | 87.5 | |
| AB-MILBackbone=Frozen IN-1K ResNet-502023.03 | 77.1 | 86.9 | |
| Max-poolingBackbone=Frozen IN-1K ResNet-502023.03 | 64.4 | 82.6 | |
| KNN (Mean)Backbone=Frozen IN-1K ResNet-502023.03 | 63.3 | 74.9 | |
| KNN (Max)Backbone=Frozen IN-1K ResNet-502023.03 | 52.4 | 63.9 |