Skin Lesion Classification on ISIC 2018 (test)
97.7AUCFedAvg
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FedAvgLabeled Data Ratio=100%2023.04 | 97.7 | 80 | 80.9 | 88.9 | 74.3 | |
| MWNL2023.07 | 96.79 | 76.94 | 74.92 | 84.9 | 73.9 | |
| SCL2023.07 | 96.56 | 80.88 | 74.27 | 86.13 | 70.4 | |
| ECL2023.07 | 96.55 | 83.44 | 76.76 | 87.2 | 73.01 | |
| Hybrid-SC2023.07 | 96.33 | 75.84 | 74.34 | 86.3 | 73.93 | |
| FedLSM2023.04 | 96 | 69.9 | 71.3 | 84.6 | 76.3 | |
| TSC2023.07 | 95.83 | 77.77 | 74.94 | 85.94 | 73.35 | |
| BCL2023.07 | 95.61 | 71.15 | 71.57 | 84.92 | 72.87 | |
| Logit Adjust2023.07 | 95.55 | 71.78 | 70.77 | 84.15 | 71.54 | |
| LDAM-DRW2023.07 | 95.22 | 74.65 | 71.98 | 84.2 | 71.74 | |
| FedProxModel Aggregation Strategy=task-dependent2023.04 | 95.2 | 61.2 | 63 | 82 | 76.8 | |
| FPSL2023.04 | 95.2 | 61.3 | 63.8 | 82.5 | 72.8 | |
| FedAvgModel Aggregation Strategy=task-dependent2023.04 | 94.9 | 59.7 | 62 | 81.7 | 75.3 | |
| CE2023.07 | 94.81 | 73.62 | 70.34 | 83.89 | 69.56 | |
| MOONModel Aggregation Strategy=task-dependent2023.04 | 94.8 | 62 | 65.2 | 82.6 | 75.5 | |
| Focal Loss2023.07 | 94.76 | 76.69 | 71.38 | 84.19 | 68.78 | |
| ACPLLabeled training data percentage=20%, Backbone=DenseNet-1212021.11 | 94.36 | 72.14 | 62.23 | — | — | |
| FedAvg + FixMatchModel Aggregation Strategy=task-dependent2023.04 | 94 | 54.1 | 56.4 | 78.9 | 68.1 | |
| FSSLModel Aggregation Strategy=task-dependent2023.04 | 93.9 | 58 | 60.8 | 80.7 | 74 | |
| BBN2023.07 | 93.72 | 72.4 | 72.79 | 85.57 | 74.96 | |
| SRC-MTLabeled training data percentage=20%2021.11 | 93.58 | 71.47 | 60.68 | — | — | |
| MTLabeled training data percentage=20%2021.11 | 92.96 | 69.75 | 59.1 | — | — | |
| TELabeled training data percentage=20%2021.11 | 92.7 | 69.81 | 59.33 | — | — | |
| FedAvg2023.04 | 92.7 | 62.2 | 57.6 | 81 | 72.1 | |
| FedPU2023.04 | 92.7 | 57 | 55 | 79.6 | 69.9 | |
| FedRS2023.04 | 92.6 | 59.7 | 57.7 | 80 | 71.6 | |
| TCSELabeled training data percentage=20%2021.11 | 92.24 | 68.17 | 58.44 | — | — | |
| SS-DCGANLabeled training data percentage=20%2021.11 | 91.28 | 67.72 | 54.1 | — | — | |
| Self-trainingLabeled training data percentage=20%2021.11 | 90.58 | 67.63 | 54.51 | — | — | |
| SupervisedLabeled training data percentage=20%, Backbone=DenseNet-1212021.11 | 90.15 | 65.5 | 52.03 | — | — | |
| CLIPEvaluation Protocol=Zero-shot2025.10 | — | — | — | 14.6 | — | |
| CoMatchData Augmentation=RandAugment (RA), Method Category=Conventional SSL2025.10 | — | — | — | 93.39 | — | |
| FixMatchData Augmentation=RandAugment (RA), Method Category=Conventional SSL2025.10 | — | — | — | 93.23 | — | |
| FlexMatchData Augmentation=RandAugment (RA), Method Category=Conventional SSL2025.10 | — | — | — | 93.4 | — | |
| PromptSRCEvaluation Protocol=Prompt tuning2025.10 | — | — | — | 59.97 | — | |
| PromptSRC + Bi-CoGEvaluation Protocol=Prompt tuning2025.10 | — | — | — | 61.03 | — | |
| Self-trainingMethod Category=Conventional SSL2025.10 | — | — | — | 95.02 | — |