mlsanityData to 7 Oct 2026

LF-Net: Learning Local Features from Images

arXiv 1805.09662 ↗First posted 24 May 2018cs.CVcited by 151

Not on the map. Why?

Abstract (arXiv)

We present a novel deep architecture and a training strategy to learn a local feature pipeline from scratch, using collections of images without the need for human supervision. To do so we exploit depth and relative camera pose cues to create a virtual target that the network should achieve on one image, provided the outputs of the network for the other image. While this process is inherently non-differentiable, we show that we can optimize the network in a two-branch setup by confining it to one branch, while preserving differentiability in the other. We train our method on both indoor and outdoor datasets, with depth data from 3D sensors for the former, and depth estimates from an off-the-shelf Structure-from-Motion solution for the latter. Our models outperform the state of the art on sparse feature matching on both datasets, while running at 60+ fps for QVGA images.

Key facts. Cited by 151 papers in the map (data through 7 Oct 2026). First posted on arXiv 24 May 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 41; 2022: 34; 2023: 37; 2024: 13; 2025: 19; 2026: 7.

151 papers.

Sort:

By area:

  1. 2021DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Camerascited by 673
  2. 2021BARF: Bundle-Adjusting Neural Radiance Fieldscited by 641
  3. 2023Emergent Correspondence from Image Diffusioncited by 437
  4. 2023A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondencecited by 297
  5. 2023TAPIR: Tracking Any Point with per-frame Initialization and temporal Refinementcited by 281
  6. 2022TAP-Vid: A Benchmark for Tracking Any Point in a Videocited by 258
  7. 2021Pixel-Perfect Structure-from-Motion with Featuremetric Refinementcited by 199
  8. 2021COTR: Correspondence Transformer for Matching Across Imagescited by 182
  9. 2023ALIKED: A Lighter Keypoint and Descriptor Extraction Network via Deformable Transformationcited by 120
  10. 2022TransVPR: Transformer-based place recognition with multi-level attention aggregationcited by 102
  11. 2021BundleTrack: 6D Pose Tracking for Novel Objects without Instance or Category-Level 3D Modelscited by 87
  12. 2021PDC-Net+: Enhanced Probabilistic Dense Correspondence Networkcited by 80
38 more
  1. 2021ALIKE: Accurate and Lightweight Keypoint Detection and Descriptor Extractioncited by 77
  2. 2021Unsupervised Object-Level Representation Learning from Scene Imagescited by 72
  3. 2021Spatially Consistent Representation Learningcited by 64
  4. 2022ClusterGNN: Cluster-based Coarse-to-Fine Graph Neural Network for Efficient Feature Matchingcited by 62
  5. 2024FlowMap: High-Quality Camera Poses, Intrinsics, and Depth via Gradient Descentcited by 61
  6. 2021DiscoBox: Weakly Supervised Instance Segmentation and Semantic Correspondence from Box Supervisioncited by 59
  7. 2024OmniGlue: Generalizable Feature Matching with Foundation Model Guidancecited by 55
  8. 2021Wide-Baseline Relative Camera Pose Estimation with Directional Learningcited by 49
  9. 2022Iterative Deep Homography Estimationcited by 45
  10. 2021P2-Net: Joint Description and Detection of Local Features for Pixel and Point Matchingcited by 42
  11. 2024Deep Learning-Based Object Pose Estimation: A Comprehensive Surveycited by 39
  12. 2021Deep Lucas-Kanade Homography for Multimodal Image Alignmentcited by 38
  13. 2023Differentiable Registration of Images and LiDAR Point Clouds with VoxelPoint-to-Pixel Matchingcited by 35
  14. 2022BALF: Simple and Efficient Blur Aware Local Feature Detectorcited by 34
  15. 2022TopicFM: Robust and Interpretable Topic-Assisted Feature Matchingcited by 32
  16. 2023Doppelgangers: Learning to Disambiguate Images of Similar Structurescited by 32
  17. 2021VS-Net: Voting with Segmentation for Visual Localizationcited by 29
  18. 2023PMatch: Paired Masked Image Modeling for Dense Geometric Matchingcited by 29
  19. 2021LocalTrans: A Multiscale Local Transformer Network for Cross-Resolution Homography Estimationcited by 28
  20. 2022Decoupling Makes Weakly Supervised Local Feature Bettercited by 28
  21. 2023SFD2: Semantic-guided Feature Detection and Descriptioncited by 28
  22. 2022Self-Supervised Equivariant Learning for Oriented Keypoint Detectioncited by 26
  23. 2024Local Feature Matching Using Deep Learning: A Surveycited by 26
  24. 2022Virtual Correspondence: Humans as a Cue for Extreme-View Geometrycited by 25
  25. 2023Zero-Shot Everything Sketch-Based Image Retrieval, and in Explainable Stylecited by 25
  26. 2023A Robust and Interpretable Deep Learning Framework for Multi-modal Registration via Keypointscited by 25
  27. 2023Learning Rotation-Equivariant Features for Visual Correspondencecited by 23
  28. 2023Adaptive Spot-Guided Transformer for Consistent Local Feature Matchingcited by 22
  29. 2023Steerers: A framework for rotation equivariant keypoint descriptorscited by 21
  30. 2022Partially Does It: Towards Scene-Level FG-SBIR with Partial Inputcited by 20
  31. 2022End2End Multi-View Feature Matching with Differentiable Pose Optimizationcited by 19
  32. 2021DFM: A Performance Baseline for Deep Feature Matchingcited by 17
  33. 2021Wide-Baseline Multi-Camera Calibration using Person Re-Identificationcited by 15
  34. 2023Segment Anything Model is a Good Teacher for Local Feature Learningcited by 15
  35. 2025Dense-SfM: Structure from Motion with Dense Consistent Matchingcited by 15
  36. 2021Contextual Similarity Aggregation with Self-attention for Visual Re-rankingcited by 14
  37. 2023SACReg: Scene-Agnostic Coordinate Regression for Visual Localizationcited by 14
  38. 2023Deep Learning for Visual Localization and Mapping: A Surveycited by 14

50 most cited of 151 All 151 (JSON)