Pelvic fracture detection on Pelvic X-ray INVIS (test)
0.8688AUCPelFANet
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PelFANetArchitecture=dual-input, attention-fused framework2025.09 | 0.8688 | 81.23 | |
| DRR20Backbone=ResNet-based, Pretraining=Digitally Reconstructed Radiographs (DRR) synthetic data2025.09 | 0.8002 | 78.6 | |
| ImageNetBackbone=ResNet-based, Pretraining=ImageNet2025.09 | 0.7549 | 72.7 | |
| ImageNet+DRR20Backbone=ResNet-based, Pretraining=ImageNet and DRR2025.09 | 0.714 | 72.1 | |
| ImageNet+DRR20_FullBackbone=ResNet-based, Pretraining=ImageNet and DRR (Full)2025.09 | 0.6896 | 77.5 |