Pose Estimation
Pose Estimation is a computer vision task where the goal is to detect the position and orientation of a person or an object. Usually, this is done by predicting the location of specific keypoints like hands, head, elbows, etc. in case of Human Pose Estimation.
A common benchmark for this task is MPII Human Pose
( Image credit: Real-time 2D Multi-Person Pose Estimation on CPU: Lightweight OpenPose )
Papers
Showing 1–10 of 4228 papers
All datasetsCOCO test-devMPII Human PoseOCHumanLeeds Sports PosesCrowdPoseCOCO val2017AICCOCO (Common Objects in Context)InLocITOP front-viewJ-HMDBMPII Single Person
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | BUCTD (PETR, with generative sampling) | APL | 83.7 | — | Unverified |
| 2 | OmniPose (WASPv2) | AP | 79.5 | — | Unverified |
| 3 | MetaPrompt-SD | AP | 79 | — | Unverified |
| 4 | Hulk(Finetune, ViT-L) | AP | 78.7 | — | Unverified |
| 5 | BUCTD (PETR, with generative sampling) | AP | 77.8 | — | Unverified |
| 6 | Hulk(Finetune, ViT-B) | AP | 77.5 | — | Unverified |
| 7 | I²R-Net (1st stage:HRFormer-B) | AP | 77.3 | — | Unverified |
| 8 | PATH (Partial FT) | AP | 77.1 | — | Unverified |
| 9 | SOLIDER (swin-B) | AP | 76.6 | — | Unverified |
| 10 | PEFORMER-Xcit-dino-p8 | AP | 72.6 | — | Unverified |