nihalsid/retrieval-fuse

[ICCV21] Code for "RetrievalFuse: Neural 3D Scene Reconstruction with a Database"

29
/ 100
Experimental

This project helps researchers and engineers working with 3D data to reconstruct detailed 3D scenes from low-resolution inputs. You input either a low-resolution distance field grid or a point cloud, and the system outputs a high-resolution, precise 3D scene model. This is ideal for those who need to generate high-fidelity 3D representations from limited or coarse sensor data.

No commits in the last 6 months.

Use this if you need to create accurate, detailed 3D models of objects or environments from low-resolution 3D scans or point clouds, leveraging existing scene geometry databases to enhance detail.

Not ideal if you are looking for a simple, out-of-the-box software application or if your primary goal is real-time 3D reconstruction without access to substantial 3D scene databases.

3D-reconstruction computer-vision scene-understanding point-cloud-processing digital-twin-creation
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 8 / 25
Community 12 / 25

How are scores calculated?

Stars

82

Forks

8

Language

Python

License

Last pushed

Oct 21, 2021

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/nihalsid/retrieval-fuse"

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