MegEngine/FINet

This is the official MegEngine implementation of FINet: Dual Branches Feature Interaction for Partial-to-Partial Point Cloud Registration, AAAI 2022

27
/ 100
Experimental

This project helps engineers and researchers accurately align 3D shapes or objects from partial scans. It takes in two incomplete 3D point clouds—one source and one reference—and outputs the precise rotation and translation needed to perfectly overlap their common areas. This is useful for anyone working with 3D scanning, object reconstruction, or robotic perception.

No commits in the last 6 months.

Use this if you need to precisely align two incomplete 3D scans of the same object, especially when only parts of the objects are visible or scanned.

Not ideal if you are working with full 3D models rather than partial scans, or if your primary need is object detection or classification instead of alignment.

3D-scanning robot-perception computer-vision object-reconstruction metrology
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 4 / 25

How are scores calculated?

Stars

29

Forks

1

Language

Python

License

MIT

Last pushed

Sep 17, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/MegEngine/FINet"

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