Animal Pose Estimation on AP-10K (val)
80.4APViTPose++
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ViTPose++Training Dataset=Mixed Datasets (387K), Backbone=MAE/ViT-Large, Resolution=256x1922025.03 | 80.4 | 97.6 | 88.5 | 52.7 | 80.8 | — | |
| SuperAnimalTraining Dataset=Quadruped-80K, Backbone=HRNet-W32, Resolution=256x2562025.03 | 80.1 | — | — | — | — | — | |
| X-Pose-VTraining Dataset=UniKPT(226K), Backbone=Swin-Large, Resolution=256x2562025.03 | 79 | 95.7 | 86.8 | 57 | 79.6 | — | |
| PPAPTraining Dataset=AP-10K, Backbone=CLIP/ViT-Large, Resolution=256x2562025.03 | 78.6 | 97.1 | 86.1 | 59.3 | 78.9 | 81.7 | |
| CLAMPTraining Dataset=AP-10K, Backbone=CLIP/ViT-Large, Resolution=256x2562025.03 | 77.8 | 96.8 | 85 | 58.7 | 78.1 | 81 | |
| PPAPTraining Dataset=AP-10K, Backbone=CLIP/ViT-Base, Resolution=256x2562025.03 | 75.4 | 95.8 | — | 53.4 | 75.9 | 78.7 | |
| ViTPose++Training Dataset=Mixed Datasets(387K), Backbone=MAE/ViT-Base, Resolution=256x1922025.03 | 74.5 | 94.9 | 82.2 | 46.8 | 75 | 70 | |
| HRFormerTraining Dataset=AP-10K, Backbone=HRFormer-B, Resolution=256x2562025.03 | 74.5 | 95.9 | 81.6 | 55.8 | 74.9 | 77.6 | |
| HRNetTraining Dataset=AP-10K, Backbone=HRNet-W48, Resolution=256x2562025.03 | 74.4 | 95.9 | 80.7 | 58.9 | 74.8 | — | |
| CLAMPTraining Dataset=AP-10K, Backbone=CLIP/ViT-Base, Resolution=256x2562025.03 | 74.3 | 95.8 | 81.4 | 47.6 | 74.9 | 77.5 | |
| TransPoseTraining Dataset=AP-10K, Backbone=HRNet-W48, Resolution=256x2562025.03 | 74.2 | 95.7 | 80.8 | 56.2 | 74.6 | 77 | |
| HRNetTraining Dataset=AP-10K, Backbone=HRNet-W32, Resolution=256x2562025.03 | 73.8 | 95.8 | 80.3 | 52.3 | 74.2 | 76.9 | |
| X-Pose-VTraining Dataset=UniKPT(226K), Backbone=Swin-Tiny, Resolution=256x2562025.03 | 73.6 | 91.9 | 80.6 | 47.2 | 74.2 | — | |
| PPAPTraining Dataset=AP-10K, Backbone=CLIP/ResNet-50, Resolution=256x2562025.03 | 73.4 | 94.1 | 79.9 | 57.1 | 73.8 | 76.5 | |
| CLAMPTraining Dataset=AP-10K, Backbone=CLIP/ResNet-50, Resolution=256x2562025.03 | 72.9 | 95.4 | 79.4 | 43.2 | 73.2 | 76.3 | |
| TokenPoseTraining Dataset=AP-10K, Backbone=HRNet-W48, Resolution=256x2562025.03 | 72.7 | 94.3 | 79.2 | 60.6 | 72.7 | 75.7 | |
| SimpleBaselineTraining Dataset=AP-10K, Backbone=ResNet-50, Resolution=256x2562025.03 | 70.2 | 94.2 | 76 | 45.5 | 70.4 | 73.5 |