zju3dv/OnePose

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1010 stars · 85 forks observed · 2026-08-28

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

OnePose is the official PyTorch implementation of the CVPR 2022 paper 'One-Shot Object Pose Estimation without CAD Models'. It estimates the 6D pose of novel objects from a single reference video, using SuperPoint/SuperGlue feature matching and COLMAP structure-from-motion instead of CAD models.

Use cases

• estimate 6d pose of an object from a single reference video
• object pose estimation without cad models
• one-shot object pose estimation for ar applications
• run 2d-3d feature matching for pose estimation
• reproduce cvpr 2022 onepose benchmark results
• capture custom object data for pose estimation training

When to choose

• you need one-shot 6D pose estimation of novel objects without CAD models
• you want to reproduce or build on the OnePose CVPR 2022 results
• you have a reference video of an object and want its pose in new scenes

When to avoid

• you need real-time pose estimation on mobile or embedded devices
• you require CAD-model-based pose estimation pipelines
• you need a maintained production-ready library rather than research code

Facets

library · maturity maintenance

computer-vision machine-learning image-processing deep-learning computer-vision machine-learning robotics python pose-estimation one-shot-learning feature-matching structure-from-motion cvpr-2022 augmented-reality 3d-vision research linux macos gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="zju3dv/OnePose")