slide-level tumor localization on Camelyon 16 (test)
88.5FROCInception (V3) ensemble
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Inception (V3) ensembleEnsemble=3 models, Input magnification=40X2017.03 | 88.5 | 92.4 | 97.7 | |
| Inception (V3)Input magnification=40X, Model size=20M parameters2017.03 | 87.3 | 91.1 | 96.7 | |
| Inception (V3) smallInput magnification=40X, Model size=300K parameters2017.03 | 86.4 | 92.4 | 97.1 | |
| Inception (V3) multi-scale smallInput magnification=40X+20X, Model size=300K parameters2017.03 | 85.9 | 92.9 | 97 | |
| Inception (V3)Input magnification=40X, Model size=20M parameters, Pre-training=ImageNet2017.03 | 85.5 | 91.1 | 97.5 | |
| Inception (V3) smallInput magnification=20X, Model size=300K parameters2017.03 | 85.5 | 91.1 | 98.6 | |
| P4M-DenseNetData regime=100%, Number of parameters=123k2018.06 | 84 | — | — | |
| Inception (V3) multi-scale smallInput magnification=40X+10X, Model size=300K parameters2017.03 | 82.2 | 87.6 | 98.6 | |
| DenseNetData regime=100%, Number of parameters=126k2018.06 | 81.7 | — | — | |
| P4M-DenseNetData regime=50%, Number of parameters=123k2018.06 | 81.5 | — | — | |
| Wang et al.Data regime=100%, Resolution=40x2018.06 | 80.7 | — | — | |
| Camelyon16 winner2017.03 | 80.7 | 82.7 | 99.4 | |
| DenseNetData regime=50%, Number of parameters=126k2018.06 | 80 | — | — | |
| Liu et al.Data regime=100%, Augmentation/Ensemble=10x2018.06 | 79.3 | — | — | |
| Inception (V3) smallInput magnification=10X, Model size=300K parameters2017.03 | 79.3 | 84.9 | 96.5 | |
| Pathologist2018.06 | 73.3 | — | — | |
| Pathologist2017.03 | 73.3 | 73.3 | 96.6 | |
| P4M-DenseNetData regime=25%, Number of parameters=123k2018.06 | 72.6 | — | — | |
| DenseNetData regime=25%, Number of parameters=126k2018.06 | 71 | — | — | |
| P4M-DenseNetData regime=12.5%, Number of parameters=123k2018.06 | 60.7 | — | — | |
| DenseNetData regime=12.5%, Number of parameters=126k2018.06 | 55.4 | — | — | |
| Fully-supervisedScale=Single2020.11 | 52.54 | — | — | |
| DSMIL-LCScale=Multiple2020.11 | 43.71 | — | — | |
| DSMILScale=Single2020.11 | 42.96 | — | — | |
| MS-ABMILScale=Multiple2020.11 | 41.91 | — | — | |
| ABMILScale=Single2020.11 | 40.56 | — | — | |
| Max-poolingScale=Single2020.11 | 33.13 | — | — | |
| MILRNNScale=Single2020.11 | 30.48 | — | — | |
| MS-MILRNNScale=Multiple2020.11 | 27.91 | — | — | |
| Mean-poolingScale=Single2020.11 | 11.62 | — | — |