Multiple-Instance Learning on Human dataset
92.29AccuracyFPath
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FPathPooling strategy=Adaptive pool2023.08 | 92.29 | 92.18 | 90.93 | 92.63 | |
| MSA-RPPositional embedding=Relative2023.08 | 92.18 | 92.13 | 90.85 | 92.18 | |
| DSMIL2023.08 | 91.76 | 91.66 | 90.3 | 91.7 | |
| FPathPooling strategy=Max pool2023.08 | 91.44 | 91.47 | 90.01 | 91.91 | |
| MSA-LPPositional embedding=Learnable2023.08 | 90.97 | 90.87 | 89.45 | 91.09 | |
| MSA2023.08 | 90.82 | 90.82 | 89.21 | 91 | |
| FPathPooling strategy=Soft pool2023.08 | 90.47 | 90.23 | 88.83 | 90.56 | |
| AB-MIL2023.08 | 90.11 | 90.05 | 88.38 | 90.5 | |
| MSA-SPPositional embedding=2D sine-cosine2023.08 | 89.15 | 89.05 | 87.48 | 89.48 | |
| Transmil2023.08 | 88.24 | 88.13 | 86.22 | 88.21 |