# magicleap/Atlas

Atlas: End-to-End 3D Scene Reconstruction from Posed Images

Repository: https://github.com/magicleap/Atlas
Canonical: https://ross.abutalabs.com/products/magicleap-atlas
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2022-04-06T16:02:35+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2242, "days_push": 1610, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1857, forks 216 (observed 2026-08-28T04:05:45.026325+00:00)

## What it is
Atlas is a PyTorch-based deep learning model from Magic Leap that performs end-to-end 3D scene reconstruction from posed RGB images, producing TSDF volumes and meshes. It includes training, evaluation, and inference code with Docker support and pretrained models.

## Use cases
- reconstruct 3D scenes from posed images
- generate 3D meshes from RGB photos
- run TSDF fusion with a neural network
- evaluate 3D reconstruction on ScanNet
- research volumetric scene reconstruction
- build 3D models from camera poses

## When to choose
- you have posed images and need dense 3D reconstruction
- you want a learned alternative to classical TSDF fusion
- you need a research baseline for end-to-end scene reconstruction

## When to avoid
- you lack GPU hardware or posed camera data
- you need actively maintained production software
- you need reconstruction from unposed images without running COLMAP first

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, computer-vision, image-processing, graphics, simulation
- domain: computer-vision, deep-learning, machine-learning, graphics, artificial-intelligence
- platform: python
- tags: 3d-reconstruction, scene-reconstruction, tsdf-fusion, pytorch, depth-estimation, research-code, posed-images, linux, docker, gpu

## Member repositories
- magicleap/Atlas (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:45.026325+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:16:37.052567+00:00, confidence not recorded.
  - readme: https://github.com/magicleap/Atlas (fetched 2026-08-28T04:05:45.026325+00:00, sha ce088bcdb61e)
- Data as of 2026-08-30T08:39:29.467469+00:00.