Ending2015a/liepose-diffusion

[CVPR 2024] Confronting Ambiguity in 6D Object Pose Estimation via Score-Based Diffusion on SE(3)

27
/ 100
Experimental

This project helps industrial automation engineers or robotics developers accurately determine the 3D position and orientation (6D pose) of objects from a single camera image, even when objects are symmetrical or partially hidden. You provide an RGB image of an object, and it outputs precise rotation and translation data. This is ideal for tasks requiring robust object manipulation or scene understanding in complex environments.

No commits in the last 6 months.

Use this if you need to precisely track or grasp objects with symmetries or occlusions using only a standard camera, and existing methods struggle with ambiguous poses.

Not ideal if your application does not involve object pose estimation, or if you are working with non-visual data.

robotics computer-vision industrial-automation object-tracking 3D-reconstruction
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 3 / 25

How are scores calculated?

Stars

42

Forks

1

Language

C++

License

MIT

Last pushed

Sep 07, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/Ending2015a/liepose-diffusion"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.