# naver/mast3r

Grounding Image Matching in 3D with MASt3R

Repository: https://github.com/naver/mast3r
Canonical: https://ross.abutalabs.com/products/mast3r
Language: Python
License: NOASSERTION
License Family: other
Last push: 2025-06-30T07:07:35+00:00

## Health v2 (maintenance only)
Score: 37/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 29, release rhythm 35, longevity 57
- inputs: {"age_days": 810, "days_push": 429, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3088, forks 280 (observed 2026-08-28T04:07:42.199479+00:00)

## What it is
MASt3R is the official PyTorch implementation of 'Grounding Image Matching in 3D with MASt3R' (ECCV 2024), a model that performs dense 3D reconstruction and pixel-accurate image matching from unconstrained image pairs. It builds on DUSt3R and includes tooling for structure-from-motion (MASt3R-SfM), visual localization, and an interactive demo.

## Use cases
- match keypoints between two images with 3D grounding
- reconstruct 3D point clouds from image pairs
- run structure-from-motion on unordered photo collections
- visual localization against a known 3D scene
- try a demo of 3D reconstruction from photos
- train or fine-tune a 3D matching model on custom datasets

## When to choose
- you need dense correspondence or 3D point clouds from sparse, unconstrained image sets
- you want a research-grade SfM or visual-localization pipeline built on DUSt3R
- you have a GPU and can work in Python/PyTorch

## When to avoid
- you need a permissively licensed library for commercial products (CC BY-NC-SA 4.0, non-commercial only)
- you need a lightweight CPU-only matcher for production pipelines
- you want a turnkey photogrammetry app rather than a research codebase

## Facets
- artifact type: library
- maturity: active
- function: computer-vision, image-processing, machine-learning, deep-learning
- domain: computer-vision, machine-learning, deep-learning
- platform: python
- tags: 3d-reconstruction, image-matching, structure-from-motion, visual-localization, stereo, point-clouds, pytorch, non-commercial-license, research, linux, gpu, docker

## Member repositories
- naver/mast3r (main) score 37

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:42.199479+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 2026-08-30T07:27:30.834348+00:00, confidence not recorded.
  - readme: https://github.com/naver/mast3r (fetched 2026-08-28T04:07:42.199479+00:00, sha 0a80be257ca6)
- Data as of 2026-08-30T08:39:29.467469+00:00.