2D Human Pose Estimation On Human Art
Metrics
AP
AP (gt bbox)
Results
Performance results of various models on this benchmark
Model Name | AP | AP (gt bbox) | Paper Title | Repository |
|---|---|---|---|---|
| HRNet-w48 | 0.417 | 0.769 | Deep High-Resolution Representation Learning for Human Pose Estimation | - |
| RTMPose-s | 0.311 | - | RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose | - |
| HRNet-w32 | 0.399 | 0.754 | Deep High-Resolution Representation Learning for Human Pose Estimation | - |
| ED-Pose (R50) | 0.723 | / | Explicit Box Detection Unifies End-to-End Multi-Person Pose Estimation | - |
| ViTPose-h | 0.468 | 0.800 | ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation | - |
| ViTPose-s | 0.381 | 0.738 | ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation | - |
| ViTPose-l | 0.459 | 0.789 | ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation | - |
| UniPose | 0.759 | - | X-Pose: Detecting Any Keypoints | - |
| ViTpose-b | 0.410 | 0.759 | ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation | - |
| RTMPose-l | - | 0.753 | RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose | - |
0 of 10 row(s) selected.