Skin Lesion Classification on ISIC 3-way 3-shot 2018 (test)
75.82AccuracyHybrid distil
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Hybrid distilBackbone=ResNet-122021.10 | 75.82 | 73.34 | |
| MoCo_v1 (baseline)Backbone=ResNet-122021.10 | 74.92 | 69.5 | |
| MoCo_v2Backbone=ResNet-122021.10 | 70.84 | 64.77 | |
| Hybrid distilBackbone=4 conv blocks2021.10 | 69.14 | 62.56 | |
| MoCo_v1 (baseline)Backbone=4 conv blocks2021.10 | 67.48 | 59.77 | |
| SRL-distilBackbone=ResNet-122021.10 | 64.92 | 59.88 | |
| SimCLRBackbone=ResNet-122021.10 | 63.82 | 57.55 | |
| SRL-simpleBackbone=ResNet-122021.10 | 61.31 | 57.65 | |
| SimCLRBackbone=4 conv blocks2021.10 | 60.78 | 55.32 | |
| MoCo_v2Backbone=4 conv blocks2021.10 | 59.71 | 52.68 | |
| SRL-distilBackbone=4 conv blocks2021.10 | 58.2 | 52.24 | |
| SRL-simpleBackbone=4 conv blocks2021.10 | 57.5 | 45.56 | |
| DAMLBackbone=ResNet-122021.10 | 55.57 | 49.01 | |
| MAMLBackbone=ResNet-122021.10 | 55.55 | 49.19 | |
| DAMLBackbone=4 conv blocks2021.10 | 53.91 | 44.8 | |
| MAMLBackbone=4 conv blocks2021.10 | 51.56 | 45.38 | |
| UMTRABackbone=ResNet-122021.10 | 51.29 | 45.91 | |
| RelationNetBackbone=ResNet-122021.10 | 47.29 | 43.37 | |
| UMTRABackbone=4 conv blocks2021.10 | 45.76 | 39.07 | |
| CACTUS-ProtoNetsBackbone=ResNet-122021.10 | 45 | 39.69 | |
| CACTUS-MAMLBackbone=ResNet-122021.10 | 44.44 | 39.94 | |
| RelationNetBackbone=4 conv blocks2021.10 | 42.63 | 38.49 | |
| CACTUS-MAMLBackbone=4 conv blocks2021.10 | 42.2 | 38.7 | |
| Training from scratchBackbone=ResNet-122021.10 | 39.76 | 35.6 | |
| Training from scratchBackbone=4 conv blocks2021.10 | 38.95 | 33.24 | |
| ProtoNetsBackbone=ResNet-122021.10 | 38.59 | 33.11 | |
| ProtoNetsBackbone=4 conv blocks2021.10 | 36.38 | 31.65 | |
| CACTUS-ProtoNetsBackbone=4 conv blocks2021.10 | 35.08 | 30.95 |