Human Pose Estimation on PoseTrack 2017 (val)
93.3HeadTAR-ViTPose*
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| TAR-ViTPose*Backbone=ViT-H, Bounding Box Source=Ground-truth2026.03 | 93.3 | 94.4 | 90.6 | 87 | 86.5 | 90 | 88.9 | — | 90.3 | |
| TAR-ViTPose*Backbone=ViT-B, Bounding Box Source=Ground-truth2026.03 | 92.3 | 92.6 | 88.2 | 83.3 | 85.9 | 87.4 | 82.2 | — | 87.7 | |
| Poseidon*Backbone=ViT-H, Bounding Box Source=Ground-truth2026.03 | 91.9 | 93.1 | 88.6 | 85.7 | 88.2 | 89.6 | 84.4 | — | 88.9 | |
| GNN-based pose dynamics learningBackbone=HRNet, Detector=Faster R-CNN, Filtering=None (pure pose estimation)2021.06 | 90.9 | 90.7 | 86 | 79.2 | 83.8 | 82.7 | 78 | 84.9 | — | |
| GLSMambaBackbone=ViT-H, Bounding Box Source=Ground-truth or unspecified2026.03 | 90.7 | 92.1 | 89.2 | 85.3 | 87 | 88.4 | 82.4 | — | 88 | |
| TDMI-ST2023.03 | 90.6 | 90.7 | 87.2 | 81.5 | 85.2 | 84.5 | 78.7 | 85.9 | — | |
| TDMIBackbone=HRNet-W48, Bounding Box Source=Ground-truth or unspecified2026.03 | 90.6 | 91 | 87.2 | 81.5 | 85.2 | 84.5 | 78.7 | — | 85.9 | |
| TAR-ViTPoseBackbone=ViT-H, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 90.5 | 91.9 | 88.4 | 83.8 | 84.9 | 87.3 | 80.2 | — | 86.8 | |
| DCPose + HANetBase method=DCPose2022.11 | 90 | 90 | 85 | 78.8 | 83.1 | 82.1 | 77.1 | 84.2 | — | |
| TDMI2023.03 | 90 | 91.1 | 87.1 | 81.4 | 85.2 | 84.5 | 78.5 | 85.7 | — | |
| DSTABackbone=HRNet-W48, Approach=regression-based2024.03 | 89.8 | 90.8 | 86.2 | 79.3 | 85.2 | 82.2 | 75.9 | 84.6 | — | |
| DSTABackbone=HRNet-W48, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 89.8 | 90.8 | 86.2 | 79.3 | 85.2 | 82.2 | 75.9 | — | 84.6 | |
| FAMI-Pose2022.03 | 89.6 | 90.1 | 86.3 | 80 | 84.6 | 83.4 | 77 | 84.8 | — | |
| FAMI-Pose2023.03 | 89.6 | 90.1 | 86.3 | 80 | 84.6 | 83.4 | 77 | 84.8 | — | |
| FAMIPoseBackbone=HRNet-W48, Approach=heatmap-based2024.03 | 89.6 | 90.1 | 86.3 | 80 | 84.6 | 83.4 | 77 | 84.8 | — | |
| FAMI-PoseBackbone=HRNet-W48, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 89.6 | 90.1 | 86.3 | 80 | 84.6 | 83.4 | 77 | — | 84.8 | |
| CombDetFiltering=None (pure pose estimation)2021.06 | 89.4 | 89.7 | 85.5 | 79.5 | 82.4 | 80.8 | 76.4 | 83.8 | — | |
| DetTrack2023.03 | 89.4 | 89.7 | 85.5 | 79.5 | 82.4 | 80.8 | 76.4 | 83.8 | — | |
| DetTrackBackbone=HRNet-W48, Approach=heatmap-based2024.03 | 89.4 | 89.7 | 85.5 | 79.5 | 82.4 | 80.8 | 76.4 | 83.8 | — | |
| DetTrackBackbone=HRNet-W48, Bounding Box Source=Ground-truth or unspecified2026.03 | 89.4 | 89.7 | 85.5 | 79.5 | 82.4 | 80.8 | 76.4 | — | 83.8 | |
| DSTABackbone=ViT-H, Approach=regression-based2024.03 | 89.3 | 90.6 | 87.3 | 82.6 | 84.5 | 85.1 | 77.8 | 85.6 | — | |
| DSTABackbone=ViT-H, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 89.3 | 90.6 | 87.3 | 82.6 | 84.5 | 85.1 | 77.8 | — | 85.6 | |
| CM-PoseBackbone=ViT, Bounding Box Source=Ground-truth or unspecified2026.03 | 89.2 | 92 | 89 | 85.6 | 88.6 | 87.2 | 81.1 | — | 87.5 | |
| DiffPoseBackbone=ViT, Bounding Box Source=Ground-truth or unspecified2026.03 | 89 | 91.2 | 87.4 | 83.5 | 85.5 | 87.2 | 80.2 | — | 86.4 | |
| Dynamic-GNN2022.03 | 88.4 | 88.4 | 82 | 74.5 | 79.1 | 78.3 | 73.1 | 81.1 | — | |
| Dynamic-GNN2023.03 | 88.4 | 88.4 | 82 | 74.5 | 79.1 | 78.3 | 73.1 | 81.1 | — | |
| Dyn.-GNNBackbone=HRNet-W48, Approach=heatmap-based2024.03 | 88.4 | 88.4 | 82 | 74.5 | 79.1 | 78.3 | 73.1 | 81.1 | — | |
| Dyn.-GNNBackbone=HRNet-W48, Bounding Box Source=Ground-truth or unspecified2026.03 | 88.4 | 88.4 | 82 | 74.5 | 79.1 | 78.3 | 73.1 | — | 81.1 | |
| DSTABackbone=ResNet-152, Approach=regression-based2024.03 | 88.3 | 88.1 | 83.3 | 76 | 82.5 | 81.1 | 70 | 81.8 | — | |
| PAVE-Net†Backbone=Swin-L, Bounding Box Source=Predicted (Faster R-CNN), End-to-end=true2026.03 | 88.2 | 89.1 | 81.7 | 74.8 | 81.6 | 78.5 | 71.8 | — | 81.3 | |
| DCPose2022.03 | 88 | 88.7 | 84.1 | 78.4 | 83 | 81.4 | 74.2 | 82.8 | — | |
| DCPose2022.11 | 88 | 88.7 | 84.1 | 78.4 | 83 | 81.4 | 74.2 | 82.8 | — | |
| DCPose2023.03 | 88 | 88.7 | 84.1 | 78.4 | 83 | 81.4 | 74.2 | 82.8 | — | |
| DCPoseBackbone=HRNet-W48, Approach=heatmap-based2024.03 | 88 | 88.7 | 84.1 | 78.4 | 83 | 81.4 | 74.2 | 82.8 | — | |
| DCPoseBackbone=HRNet-W48, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 88 | 88.7 | 84.1 | 78.4 | 83 | 81.4 | 74.2 | — | 82.8 | |
| TAR-ViTPoseBackbone=ViT-B, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 88 | 89.2 | 85 | 80.3 | 82.9 | 84.2 | 77.3 | — | 84 | |
| TAR-ViTPose*Backbone=ViT-S, Bounding Box Source=Ground-truth2026.03 | 88 | 88.9 | 84.2 | 78.4 | 82.6 | 81.5 | 76.9 | — | 83.2 | |
| TAR-ViTPoseBackbone=ViT-S, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 87.1 | 88 | 83.2 | 77.8 | 80.4 | 81 | 74.2 | — | 81.9 | |
| CorrTrack2023.03 | 86.1 | 87 | 83.4 | 76.4 | 77.3 | 79.2 | 73.3 | 80.8 | — | |
| MDPN2022.03 | 85.2 | 88.5 | 83.9 | 77.5 | 79 | 77 | 71.4 | 80.7 | — | |
| MDPN2023.03 | 85.2 | 88.5 | 83.9 | 77.5 | 79 | 77 | 71.4 | 80.7 | — | |
| MDPNBackbone=ResNet-152, Approach=heatmap-based2024.03 | 85.2 | 88.5 | 83.9 | 77.5 | 79 | 77 | 71.4 | 80.7 | — | |
| STEmbedding2022.03 | 83.8 | 81.6 | 77.1 | 70 | 77.4 | 74.5 | 70.8 | 77 | — | |
| STEmbedding2023.03 | 83.8 | 81.6 | 77.1 | 70 | 77.4 | 74.5 | 70.8 | 77 | — | |
| STEmbed.Backbone=ResNet-152, Approach=heatmap-based2024.03 | 83.8 | 81.6 | 77.1 | 70 | 77.4 | 74.5 | 70.8 | 77 | — | |
| STEmbedBackbone=ResNet-152, Bounding Box Source=Ground-truth or unspecified2026.03 | 83.8 | 81.6 | 77.1 | 70 | 77.4 | 74.5 | 70.8 | — | 77 | |
| HRNet2022.03 | 82.1 | 83.6 | 80.4 | 73.3 | 75.5 | 75.3 | 68.5 | 77.3 | — | |
| HRNet2022.11 | 82.1 | 83.6 | 80.4 | 73.3 | 75.5 | 75.3 | 68.5 | 77.3 | — | |
| HRNet2023.03 | 82.1 | 83.6 | 80.4 | 73.3 | 75.5 | 75.3 | 68.5 | 77.3 | — | |
| HRNetBackbone=HRNet-W48, Approach=heatmap-based2024.03 | 82.1 | 83.6 | 80.4 | 73.3 | 75.5 | 75.3 | 68.5 | 77.3 | — | |
| HRNetBackbone=HRNet-W48, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 82.1 | 83.6 | 80.4 | 73.3 | 75.5 | 75.3 | 68.5 | — | 77.3 | |
| Simplebackbone=ResNet-1522022.03 | 81.7 | 83.4 | 80 | 72.4 | 75.3 | 74.8 | 67.1 | 76.7 | — | |
| SimpleBackbone=ResNet-1522023.03 | 81.7 | 83.4 | 80 | 72.4 | 75.3 | 74.8 | 67.1 | 76.7 | — | |
| SimBase.Backbone=ResNet-152, Approach=heatmap-based2024.03 | 81.7 | 83.4 | 80 | 72.4 | 75.3 | 74.8 | 67.1 | 76.7 | — | |
| SimBase.Backbone=ResNet-152, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 81.7 | 83.4 | 80 | 72.4 | 75.3 | 74.8 | 67.1 | — | 76.7 | |
| PoseWarperFiltering=None (pure pose estimation)2021.06 | 81.4 | 88.3 | 83.9 | 78 | 82.4 | 80.5 | 73.6 | 81.2 | — | |
| PoseWarper2022.03 | 81.4 | 88.3 | 83.9 | 78 | 82.4 | 80.5 | 73.6 | 81.2 | — | |
| PoseWarper2022.11 | 81.4 | 88.3 | 83.9 | 78 | 82.4 | 80.5 | 73.6 | 81.2 | — | |
| PoseWarper2023.03 | 81.4 | 88.3 | 83.9 | 78 | 82.4 | 80.5 | 73.6 | 81.2 | — | |
| PoseWarp.Backbone=HRNet-W48, Approach=heatmap-based2024.03 | 81.4 | 88.3 | 83.9 | 78 | 82.4 | 80.5 | 73.6 | 81.2 | — | |
| PoseWarp.Backbone=HRNet-W48, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 81.4 | 88.3 | 83.9 | 78 | 82.4 | 80.5 | 73.6 | — | 81.2 | |
| FastPose2022.03 | 80 | 80.3 | 69.5 | 59.1 | 71.4 | 67.5 | 59.4 | 70.3 | — | |
| FastPose2023.03 | 80 | 80.3 | 69.5 | 59.1 | 71.4 | 67.5 | 59.4 | 70.3 | — | |
| FastPoseBackbone=ResNet-101, Approach=heatmap-based2024.03 | 80 | 80.3 | 68.5 | 59.1 | 71.4 | 67.5 | 59.4 | 70.3 | — | |
| Simplebackbone=ResNet-502022.03 | 79.1 | 80.5 | 75.5 | 66 | 70.8 | 70 | 61.7 | 72.4 | — | |
| SimpleBackbone=ResNet-502023.03 | 79.1 | 80.5 | 75.5 | 66 | 70.8 | 70 | 61.7 | 72.4 | — | |
| JointFlow2022.03 | 69.3 | — | — | — | — | — | — | 69.3 | — | |
| PoseTracker2022.03 | 67.5 | 70.2 | 62 | 51.7 | 60.7 | 58.7 | 49.8 | 60.6 | — | |
| PoseTracker2023.03 | 67.5 | 70.2 | 62 | 51.7 | 60.7 | 58.7 | 49.8 | 60.6 | — | |
| PoseTrackBackbone=ResNet-101, Approach=heatmap-based2024.03 | 67.5 | 70.2 | 62 | 51.7 | 60.7 | 58.7 | 49.8 | 60.6 | — | |
| PoseTrackBackbone=ResNet-101, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 67.5 | 70.2 | 62 | 51.7 | 60.7 | 58.7 | 49.8 | — | 60.6 | |
| PoseFlow2022.03 | 66.7 | 73.3 | 68.3 | 61.1 | 67.5 | 67 | 61.3 | 66.5 | — | |
| PoseFlow2023.03 | 66.7 | 73.3 | 68.3 | 61.1 | 67.5 | 67 | 61.3 | 66.5 | — | |
| PoseFlowBackbone=ResNet-101, Approach=heatmap-based2024.03 | 66.7 | 73.3 | 68.3 | 61.1 | 67.5 | 67 | 61.3 | 66.5 | — | |
| PoseFlowBackbone=ResNet-152, Bounding Box Source=Predicted (Faster R-CNN)2026.03 | 66.7 | 73.3 | 68.3 | 61.1 | 67.5 | 67 | 61.3 | — | 66.5 | |
| DCPoseBackbone=HRNet-W48, #Params=65.2M, FPS(↑)=128, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 82.8 | |
| DSTABackbone=ResNet-50, Method Category=video-based, Input Resolution=256 x 1922024.03 | — | — | — | — | — | — | — | — | 79.7 | |
| DSTABackbone=HRNet-W48, Method Category=video-based, Input Resolution=384 x 2882024.03 | — | — | — | — | — | — | — | — | 84.6 | |
| DSTABackbone=ViT-H, Method Category=video-based, Input Resolution=256 x 1922024.03 | — | — | — | — | — | — | — | — | 85.6 | |
| DSTABackbone=HRNet-W48, #Params=63.7M, FPS(↑)=156, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 83.4 | |
| DSTABackbone=ViT-H, #Params=422.2M, FPS(↑)=25, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 84.3 | |
| JointFlow2023.03 | — | — | — | — | — | — | — | 69.3 | — | |
| PoseurBackbone=ResNet-50, Method Category=image-based, Input Resolution=256 x 1922024.03 | — | — | — | — | — | — | — | — | 74.4 | |
| PoseurBackbone=HRNet-W48, Method Category=image-based, Input Resolution=384 x 2882024.03 | — | — | — | — | — | — | — | — | 79.3 | |
| PoseurBackbone=ViT-H, Method Category=image-based, Input Resolution=256 x 1922024.03 | — | — | — | — | — | — | — | — | 81 | |
| PoseWarperBackbone=HRNet-W48, #Params=71.1M, FPS(↑)=52, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 81 | |
| RLEBackbone=ResNet-50, Method Category=image-based, Input Resolution=256 x 1922024.03 | — | — | — | — | — | — | — | — | 70.7 | |
| RLEBackbone=HRNet-W48, Method Category=image-based, Input Resolution=384 x 2882024.03 | — | — | — | — | — | — | — | — | 75.7 | |
| RLEBackbone=ViT-H, Method Category=image-based, Input Resolution=256 x 1922024.03 | — | — | — | — | — | — | — | — | 79 | |
| TAR-ViTPoseBackbone=ViT-S, #Params=35.6M, FPS(↑)=413, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 81.5 | |
| TAR-ViTPoseBackbone=ViT-B, #Params=101.1M, FPS(↑)=186, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 83.4 | |
| TAR-ViTPoseBackbone=ViT-H, #Params=672.5M, FPS(↑)=28, GPU=A6000, Batch size=16, Auxiliary frames=22026.03 | — | — | — | — | — | — | — | — | 86.3 | |
| TML++2022.03 | — | — | — | — | — | — | — | 71.5 | — | |
| TML++2023.03 | — | — | — | — | — | — | — | 71.5 | — |