Medical Image Classification on BreaKHis downsampling quarter (x4↓)
86.81AccuracyMH-pFLGB
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MH-pFLGBModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 86.81 | 83.33 | |
| pFedDFModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 81.21 | 75.34 | |
| FedDFModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 78.26 | 73.42 | |
| KT-pFLModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 78.01 | 73.25 | |
| DS-PFLModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 77.82 | 72.58 | |
| FedRepModel Architecture=Unified (ResNet17)2024.06 | 77.62 | 71.82 | |
| FedMDModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 77.21 | 68.93 | |
| FedProx+FTModel Architecture=Unified (ResNet17), Fine-tuned=true2024.06 | 75.48 | 68.11 | |
| Only Local TrainingModel Architecture=Heterogeneous (ResNet17, 11, 8, 5)2024.06 | 75.38 | 68.52 | |
| SCAFFOLD+FTModel Architecture=Unified (ResNet17), Fine-tuned=true2024.06 | 75.23 | 68.72 | |
| FedAvg+FTModel Architecture=Unified (ResNet17), Fine-tuned=true2024.06 | 73.27 | 66.28 | |
| FedProxModel Architecture=Unified (ResNet17)2024.06 | 69.44 | 61.07 | |
| FedAvgModel Architecture=Unified (ResNet17)2024.06 | 68.92 | 60.31 | |
| SCAFFOLDModel Architecture=Unified (ResNet17)2024.06 | 67.25 | 59.63 | |
| APFLModel Architecture=Unified (ResNet17)2024.06 | 62.27 | 52.29 | |
| LG-FedAvgModel Architecture=Unified (ResNet17)2024.06 | 61.31 | 50.8 | |
| DittoModel Architecture=Unified (ResNet17)2024.06 | 60.65 | 50.22 |