JiawangBian/GMS-Feature-Matcher

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1114 stars · 358 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

GMS (Grid-based Motion Statistics) is a fast, robust feature correspondence algorithm for filtering feature matches between images, published at CVPR 2017 and IJCV 2020. It provides C++, Python, and Matlab implementations and has been integrated into OpenCV's xfeatures2d module.

Use cases

• filter out bad feature matches between two images
• robust feature correspondence for SLAM or visual odometry
• match ORB or SIFT features quickly
• find image correspondences for structure from motion
• remove outliers from nearest-neighbor feature matches
• integrate a fast match filter into an OpenCV pipeline

When to choose

• you need fast, robust filtering of feature matches in real-time vision systems like SLAM
• you want a lightweight alternative to RANSAC for match outlier rejection
• you are already using OpenCV and can leverage the built-in matchGMS

When to avoid

• you need deep-learning-based matchers for extreme viewpoint or appearance changes
• you need a maintained project with recent updates - the repo has not been updated since 2020
• you need matching across image collections rather than image pairs

Facets

library · maturity maintenance

computer-vision image-processing computer-vision image-processing robotics cpp python cross-platform feature-matching gms slam sfm orb sift opencv cvpr

Member repositories

For agents

markdown · JSON · MCP: product_card(name="JiawangBian/GMS-Feature-Matcher")