Human Pose Estimation on MPII (val)
94.2PCKh (Mean)ViTPose++-H
Evaluation Results
| Method | Links | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ViTPose++-HBackbone=ViT-H, Resolution=256x1922022.12 | 94.2 | — | — | — | — | — | — | 97.8 | 97.6 | 94.4 | 91.5 | 93.2 | 92.8 | 90.2 | — | — | |
| ViTPose++-LBackbone=ViT-L, Resolution=256x1922022.12 | 94 | — | — | — | — | — | — | 98 | 97.6 | 94.3 | 90.9 | 92.9 | 92.6 | 89.5 | — | — | |
| ViTPose-HBackbone=ViT-H, Resolution=256x1922022.12 | 93 | — | — | — | — | — | — | 97.7 | 97.1 | 93.2 | 89.6 | 91.8 | 91.2 | 88.1 | — | — | |
| ViTPose-LBackbone=ViT-L, Resolution=256x1922022.12 | 92.8 | — | — | — | — | — | — | 97.7 | 97.2 | 92.9 | 89.2 | 92.3 | 90.8 | 87.4 | — | — | |
| ViTPose++-BBackbone=ViT-B, Resolution=256x1922022.12 | 92.8 | — | — | — | — | — | — | 97.3 | 97.2 | 93.3 | 89.7 | 91.5 | 90.7 | 87.2 | — | — | |
| ViTPose++-SBackbone=ViT-S, Resolution=256x1922022.12 | 92.7 | — | — | — | — | — | — | 97.4 | 97.2 | 92.9 | 89 | 92.3 | 90.4 | 86.8 | — | — | |
| TransPose-H-A6Backbone=HRNet-w48, Resolution=256x2562022.12 | 92.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ASDABackbone=HRNet-w48, Resolution=256x2562022.12 | 91.4 | — | — | — | — | — | — | 97.3 | 96.5 | 91.7 | 87.9 | 90.8 | 88.2 | 84.2 | — | — | |
| ViTPose-BBackbone=ViT-B, Resolution=256x1922022.12 | 90.9 | — | — | — | — | — | — | 97 | 96.2 | 90.7 | 86.7 | 90.4 | 88.2 | 84.2 | — | — | |
| OKDHPBackbone=8-Stack Hourglass, Resolution=256x2562022.12 | 90.6 | — | — | — | — | — | — | 97.3 | 96.1 | 91.2 | 86.8 | 89.9 | 86.9 | 83.1 | — | — | |
| HRFormerBackbone=HRFormer-B, Resolution=256x2562022.12 | 90.4 | — | — | — | — | — | — | 96.8 | 96.1 | 90.4 | 85.9 | 89 | 87.3 | 84.1 | — | — | |
| HRNetBackbone=HRNet-w48, Resolution=256x2562022.12 | 90.1 | — | — | — | — | — | — | 97.1 | 95.8 | 90.7 | 85.6 | 89 | 86.8 | 82.1 | — | — | |
| CFABackbone=ResNet-101, Resolution=384x3842022.12 | 90.1 | — | — | — | — | — | — | 95.9 | 95.4 | 91 | 86.9 | 89.8 | 87.6 | 83.9 | — | — | |
| HRNetBackbone=HRNet-w32, Resolution=256x2562022.12 | 90 | — | — | — | — | — | — | 96.9 | 95.9 | 90.5 | 85.9 | 89.1 | 86.1 | 82.5 | — | — | |
| HRFormerBackbone=HRFormer-S, Resolution=256x2562022.12 | 89.9 | — | — | — | — | — | — | 97.1 | 95.8 | 90.5 | 85.9 | 88.7 | 85.7 | 82.1 | — | — | |
| SimpleBaselineBackbone=ResNet-152, Resolution=256x2562022.12 | 89 | — | — | — | — | — | — | 86.9 | 95.4 | 89.4 | 84 | 88 | 84.6 | 82.1 | — | — | |
| ViTPose-SBackbone=ViT-S, Resolution=256x1922022.12 | 88.4 | — | — | — | — | — | — | 96.4 | 94.7 | 88.1 | 83.2 | 88.4 | 84.3 | 80 | — | — | |
| Dite-HRNet-18#Params (M)=1.1, GFLOPs=0.22022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 87 | — | |
| Dite-HRNet-30#Params (M)=1.8, GFLOPs=0.42022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 87.6 | — | |
| Lite-HRNet-18#Params (M)=1.1, GFLOPs=0.22022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 86.1 | — | |
| Lite-HRNet-30#Params (M)=1.8, GFLOPs=0.42022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 87 | — | |
| MobileNetV2 1x#Params (M)=9.6, GFLOPs=1.92022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.4 | — | |
| MobileNetV3 1x#Params (M)=8.7, GFLOPs=1.82022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 84.3 | — | |
| PoseBHGFLOPS=18.5, Backbone=ViT-Base, Input Resolution=256x1922025.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.2 | |
| RLEbaseline=Regression-based2023.11 | — | 86.2 | 32.9 | 75.4 | 78.8 | 3.35 | 1.76 | — | — | — | — | — | — | — | — | — | |
| RLE + CCNetbaseline=Regression-based, CCNet=true2023.11 | — | — | — | 76.6 | — | 2.98 | 1.49 | — | — | — | — | — | — | — | — | — | |
| SBLbaseline=Heatmap-based2023.11 | — | 88.5 | 33.9 | 77.3 | 80.5 | 3.9 | 2.36 | — | — | — | — | — | — | — | — | — | |
| SBL + CCNetbaseline=Heatmap-based, CCNet=true2023.11 | — | — | — | 77.7 | — | 3.52 | 1.95 | — | — | — | — | — | — | — | — | — | |
| ShuffleNetV2 1x#Params (M)=7.6, GFLOPs=1.72022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 82.8 | — | |
| Small HRNet#Params (M)=1.3, GFLOPs=0.72022.04 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 80.2 | — | |
| UniHCPGFLOPS=18.9, Backbone=ViT-Base, Input Resolution=256x1922025.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 90.9 | |
| ViTPose++GFLOPS=18.5, Backbone=ViT-Base, Input Resolution=256x1922025.05 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 93.1 |