# cvg/LightGlue

LightGlue: Local Feature Matching at Light Speed (ICCV 2023)

Repository: https://github.com/cvg/LightGlue
Canonical: https://ross.abutalabs.com/products/lightglue
Language: Python
License: Apache-2.0
License Family: permissive
Topics: deep-learning, image-matching, pose-estimation, transformers
Last push: 2026-02-18T12:45:19+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 68, release rhythm 8, longevity 83
- inputs: {"age_days": 1165, "days_push": 196, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4728, forks 553 (observed 2026-08-28T04:08:57.613972+00:00)

## What it is
LightGlue is a deep neural network library that matches sparse local features across image pairs with high accuracy and fast inference. It takes keypoints and descriptors as input and returns corresponding point indices, using adaptive pruning of network width and depth.

## Use cases
- match keypoints between two images
- estimate camera pose from image pairs
- find corresponding points for 3D reconstruction
- run fast feature matching for visual localization
- match SuperPoint or DISK features in Python
- build image stitching or SLAM pipelines

## When to choose
- you need accurate sparse feature matching with fast inference
- you already extract keypoints with SuperPoint, DISK, ALIKED, or SIFT
- you want a research-grade matcher with pretrained weights and PyTorch integration

## When to avoid
- you need dense pixel-wise matching rather than sparse keypoints
- you cannot use a GPU or PyTorch in your environment
- you need training code, which lives in the separate glue-factory library

## Facets
- artifact type: library
- maturity: stable
- function: deep-learning, computer-vision, image-processing, machine-learning
- domain: computer-vision, deep-learning, machine-learning
- platform: python, cross-platform
- tags: feature-matching, image-matching, pose-estimation, transformers, superpoint, inference, gpu

## Member repositories
- cvg/LightGlue (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:57.613972+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 2026-08-29T18:19:02.767636+00:00, confidence not recorded.
  - readme: https://github.com/cvg/LightGlue (fetched 2026-08-28T04:08:57.613972+00:00, sha bba2e895cdfa)
- Data as of 2026-08-30T08:39:29.467469+00:00.