facebook/sam-3d-body-dinov3
AI model · https://huggingface.co/facebook/sam-3d-body-dinov3
First seen by Oriel on .
AI summary
SAM 3D Body is a promptable single-image model for full-body 3D human mesh recovery, delivering robust pose and shape estimates under occlusions, difficult poses, and challenging viewpoints.
- The encoder-decoder model accepts optional 2D keypoints and masks for user-guided inference similar to the SAM family.
- Momentum Human Rig decouples skeletal structure from surface shape while representing body, feet, and hands for improved accuracy and interpretability.
- Training annotations combine differentiable optimization, multi-view geometry, dense keypoint detection, and a data engine covering common and rare poses across varied viewpoints.
- Per-person outputs include mesh vertices, 2D and 3D keypoints, camera translation, focal length, and body, hand, and shape parameters.
- The checkpoint is available through Hugging Face, with command-line and Python inference examples, visualization tooling, and licensing under SAM License.
#human mesh recovery #computer vision #3d human pose #sam
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Likes
Total likes reported by the source.
- Hugging Face Models: 1,297, latest observation ; +371 over observed period since
Downloads (30d)
Rolling 30-day downloads. A change in this value is not new downloads during the observed period.
- Hugging Face Models: 16,078, latest observation ; +2,218 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Models | facebook/sam-3d-body-dinov3 | https://huggingface.co/facebook/sam-3d-body-dinov3 Read discussion |