facebookresearch/frankmocap

Health v2 · maintenance only

Flags: no_releases archived no_license

Adoption not part of the score

2294 stars · 387 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

FrankMocap is a single-view 3D motion capture system from Facebook AI Research that estimates 3D pose for body, hands, and whole body (body+hands) from videos and images. It packages state-of-the-art pose estimation models with demo scripts and rendering options for easy use.

Use cases

• estimate 3d body pose from a video
• capture 3d hand motion from webcam footage
• run whole-body motion capture from a single camera
• extract smpl body mesh from images
• egocentric hand tracking from head-mounted video
• generate 3d mocap output for animation or research

When to choose

• you need 3d human body or hand pose from monocular video without multi-camera rigs
• you want a research-grade pose estimator with simple demo scripts
• you need smpl-compatible 3d output for downstream graphics or analysis

When to avoid

• you need real-time production mocap with active support and updates
• you require multi-person tracking at scale or commercial licensing clarity
• you need a maintained project - the repo has seen limited recent activity

Facets

library · maturity maintenance

computer-vision machine-learning image-processing video-processing computer-vision machine-learning graphics artificial-intelligence python 3d-pose-estimation motion-capture smpl hand-pose body-pose single-view facebook-research linux macos gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="facebookresearch/frankmocap")