Efficient Loftr: Semi-dense Local Feature Matching With Sparse-like Speed

Abstract

We present a novel method for efficiently producing semi-dense matches across images. Previous detector-free matcher LoFTR has shown remarkable matching capability in handling large-viewpoint change and texture-poor scenarios but suffers from low efficiency. We revisit its design choices and derive multiple improvements for both efficiency and accuracy. One key observation is that performing the transformer over the entire feature map is redundant due to shared local information, therefore we propose an aggregated attention mechanism with adaptive token selection for efficiency. Furthermore, we find spatial variance exists in LoFTR's fine correlation module, which is adverse to matching accuracy. A novel two-stage correlation layer is proposed to achieve accurate subpixel correspondences for accuracy improvement. Our efficiency optimized model is \(\sim 2.5\times\) faster than LoFTR which can even surpass state-of-the-art efficient sparse matching pipeline SuperPoint + LightGlue. Moreo

Stats

• citations151
• S2 citations—
• github stars0
• HF likes0
• heat score16.36
• arxiv keywang2024efficient

Related papers

• Sparse-to-dense Hypercolumn Matching For Long-term Visual Localization (2019)12.99
• Local Feature Matching Using Deep Learning: A Survey (2024)18.68
• Supergf: Unifying Local And Global Features For Visual Localization (2022)0.00
• Leveraging Semantic Cues From Foundation Vision Models For Enhanced Local Feature Correspondence (2024)5.24
• Working Hard To Know Your Neighbor's Margins: Local Descriptor Learning Loss (2017)0.00
• Modality-aware Feature Matching: A Comprehensive Review Of Single- And Cross-modality Techniques (2025)0.00
• Latformer: Locality-aware Point-view Fusion Transformer For 3D Shape Recognition (2021)6.34
• Learning Local Descriptors By Optimizing The Keypoint-correspondence Criterion: Applications To Face Matching, Learning From Unlabeled Videos And 3d-shape Retrieval (2016)11.75