Human Pose Estimation on OCHuman 55 (val)
87.4APViTPose++-L
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| ViTPose++-LBackbone=ViT-L, Resolution=256x1922022.12 | 87.4 | 93.7 | 88.8 | 94.2 | |
| ViTPose++-HBackbone=ViT-H, Resolution=256x1922022.12 | 86.8 | 92.8 | 88.3 | 93.7 | |
| ViTPose++-BBackbone=ViT-B, Resolution=256x1922022.12 | 83.7 | 91.8 | 85.4 | 92.9 | |
| ViTPose++-SBackbone=ViT-S, Resolution=256x1922022.12 | 79.4 | 90.7 | 81.6 | 91.4 | |
| MIPNetBackbone=HRNet-w48, Resolution=384x2882022.12 | 74.1 | 89.7 | 81 | — | |
| ViTPose-HBackbone=ViT-H, Resolution=256x1922022.12 | 67.5 | 79.6 | 70.7 | 81.1 | |
| ViTPose-LBackbone=ViT-L, Resolution=256x1922022.12 | 65.2 | 77.2 | 69.1 | 79.8 | |
| HRNetBackbone=HRNet-w48, Resolution=384x2882022.12 | 62.1 | 76.1 | 65.9 | 78.2 | |
| HRNetBackbone=HRNet-w32, Resolution=384x2882022.12 | 60.9 | 76 | 65.1 | 78.2 | |
| ViTPose-BBackbone=ViT-B, Resolution=256x1922022.12 | 60.1 | 75.9 | 64.3 | 78 | |
| SimpleBaselineBackbone=ResNet-152, Resolution=384x2882022.12 | 58.8 | 72.7 | 63.1 | 75.7 | |
| ViTPose-SBackbone=ViT-S, Resolution=256x1922022.12 | 57.6 | 75.2 | 61.8 | 77.8 | |
| HRFormerBackbone=HRFormer-S, Resolution=384x2882022.12 | 53.1 | 73.1 | 59.6 | 76.9 | |
| HRFormerBackbone=HRFormer-B, Resolution=384x2882022.12 | 50.4 | 71.5 | 58.8 | 76.6 |