SLAM3R is a real-time RGB SLAM system that performs dense 3D reconstruction via points regression with feed-forward neural networks.

This repository contains pre-trained weights for the Image-to-Points network, which converts short video clips into dense point clouds with all frames aligned in a unified coordinate system.

Paper: https://huggingface.co/papers/2412.09401

Code: https://github.com/PKU-VCL-3DV/SLAM3R

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.