Chest Radiograph Classification on VinDr-CXR (test)
0.905AUROCConvNeXt-B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ConvNeXt-BInitialization strategy=DINOv3, Resolution=512x512, Evaluation protocol=full fine-tuning2025.10 | 0.905 | — | |
| ViT-BInitialization strategy=DINOv3, Resolution=512x512, Evaluation protocol=full fine-tuning2025.10 | 0.903 | — | |
| ViT-BInitialization strategy=DINOv3, Resolution=224x224, Evaluation protocol=full fine-tuning2025.10 | 0.902 | — | |
| ConvNeXt-BInitialization strategy=ImageNet, Resolution=512x512, Evaluation protocol=full fine-tuning2025.10 | 0.897 | 0.013 | |
| ViT-BInitialization strategy=DINOv2, Resolution=224x224, Evaluation protocol=full fine-tuning2025.10 | 0.892 | 0.11 | |
| ViT-BInitialization strategy=DINOv2, Resolution=512x512, Evaluation protocol=full fine-tuning2025.10 | 0.891 | 0.006 | |
| ViT-BInitialization strategy=ImageNet, Resolution=224x224, Evaluation protocol=full fine-tuning2025.10 | 0.883 | 0.006 | |
| ConvNeXt-BInitialization strategy=ImageNet, Resolution=224x224, Evaluation protocol=full fine-tuning2025.10 | 0.88 | 0.45 | |
| ConvNeXt-BInitialization strategy=DINOv3, Resolution=224x224, Evaluation protocol=full fine-tuning2025.10 | 0.88 | — | |
| ViT-BInitialization strategy=ImageNet, Resolution=512x512, Evaluation protocol=full fine-tuning2025.10 | 0.864 | 0.006 |