Medical Image Classification on DermaMNIST 0% Noise
97.7SensitivityDivideMix+CS
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DivideMix+CSCS Loss=True2026.04 | 97.7 | 51.8 | 74.8 | 88 | 49.4 | |
| UNICON+CSCS Loss=True2026.04 | 97.7 | 51.8 | 74.8 | 88 | 49.4 | |
| Baseline+CSCS Loss=True2026.04 | 85.2 | 75.8 | 80.5 | 89 | 59.8 | |
| GMM Filter+CSCS Loss=True2026.04 | 81.1 | 79.5 | 80.3 | 89 | 61.2 | |
| Co-teaching+CSCS Loss=True2026.04 | 79.1 | 82.1 | 80.6 | 89.6 | 62.6 | |
| GMM FilterCS Loss=False2026.04 | 75.8 | 81.8 | 78.8 | 88.9 | 60.4 | |
| Co-teachingCS Loss=False2026.04 | 75.5 | 84.6 | 80 | 89.1 | 63.2 | |
| DivideMixCS Loss=False2026.04 | 73.5 | 82.6 | 78 | 88.2 | 59.9 | |
| UNICONCS Loss=False2026.04 | 73.5 | 82.6 | 78 | 88.2 | 59.9 | |
| BaselineCS Loss=False2026.04 | 70.9 | 87.4 | 79.1 | 89.7 | 63.6 |