Research map / Video / Video segmentation and tracking
GMS: Grid-based Motion Statistics for Fast, Ultra-Robust Feature Correspondence
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| increase the number of ORB features (5,000 to 20,000 to 50,000) with GMS and no RANSAC (ablation) | GMS with ORB · wide-baseline example (Fig. 5) | match quality (qualitative) | Helped |
| use GMS grid filter instead of ratio test (threshold 0.66) | ORB matching · TUM, Strecha, VGG, Cabinet | F-measure | Helped |
| GMS on VGG with significant in-plane rotation | GMS · VGG | F-measure | Helped |
| compare GMS pose accuracy with other fast matchers | GMS · TUM pose estimation (5 deg rotation, 15 deg translation) | percentage of correct poses | Helped |
| vary the pose error threshold (ablation) | GMS · Strecha, Cabinet | pose success ratio | No clear effect |
| grouping neighboring cell-pairs (K larger) for scoring | GMS · analysis | partionability | Helped |
| finer grid with more cells at fixed 10,000 features | GMS · analysis | partionability | Hurt |
| in-plane rotation of image pairs | GMS · VGG rotation | F-measure | Mixed |
Closest papers by meaning
- UFM: A Simple Path towards Unified Dense Correspondence with Flow (NeurIPS 2025 · similarity 0.83)
- MATCHA: Towards Matching Anything (CVPR 2025 · similarity 0.83)
- GIM: Learning Generalizable Image Matcher From Internet Videos (ICLR 2024 · similarity 0.83)
- FC-GNN: Recovering Reliable and Accurate Correspondences from Interferences (CVPR 2024 · similarity 0.82)
- Improving Progressive Generation with Decomposable Flow Matching (NeurIPS 2025 · similarity 0.82)
- Adapting Dense Matching for Homography Estimation with Grid-based Acceleration (CVPR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.