Fine-Grained Visual Categorization on Dog
97.3AccuracySR-GNN
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SR-GNNInput Size=224x2242022.09 | 97.3 | 30.9 | 9.8 | |
| CAMFTransformer=Swin-224, Input Size=448x4482022.09 | 92.8 | — | — | |
| CAMFTransformer=Swin-224, Input Size=224x2242022.09 | 92.6 | — | — | |
| SwinTransformer=Swin-224, Input Size=448x4482022.09 | 92.5 | 88 | 47 | |
| TrnFGTransformer=ViT-B-16, Input Size=448x4482022.09 | 92.3 | — | — | |
| SwinTransformer=Swin-224, Input Size=224x2242022.09 | 91.8 | 88 | 15.4 | |
| ViTTransformer=ViT-B-16, Input Size=448x4482022.09 | 91.7 | 86 | 55.4 |