Multiple-Instance Learning on Rat dataset
99.56AccuracyFPath
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FPathPooling strategy=Adaptive pool2023.08 | 99.56 | 99.52 | 0.9943 | 99.49 | |
| DSMIL2023.08 | 99.51 | 99.48 | 0.9937 | 99.45 | |
| FPathPooling strategy=Soft pool2023.08 | 99.35 | 99.29 | 0.9915 | 99.24 | |
| MSA-RPPositional embedding=Relative2023.08 | 99.15 | 99.15 | 0.9896 | 99.16 | |
| FPathPooling strategy=Max pool2023.08 | 99.1 | 99.09 | 0.9888 | 99.09 | |
| Transmil2023.08 | 98.99 | 98.88 | 0.9875 | 98.78 | |
| MSA-LPPositional embedding=Learnable2023.08 | 98.79 | 98.75 | 0.9853 | 98.73 | |
| MSA2023.08 | 98.75 | 98.83 | 0.9861 | 98.92 | |
| MSA-SPPositional embedding=2D sine-cosine2023.08 | 98.51 | 98.39 | 0.981 | 98.32 | |
| AB-MIL2023.08 | 98.15 | 98.28 | 0.9793 | 98.44 |