mlsanityData to 7 Oct 2026

A case for using rotation invariant features in state of the art feature matchers

arXiv 2204.10144 ↗First posted 21 Apr 2022cs.CVcited by 27

Not on the map. Why?

Abstract (arXiv)

The aim of this paper is to demonstrate that a state of the art feature matcher (LoFTR) can be made more robust to rotations by simply replacing the backbone CNN with a steerable CNN which is equivariant to translations and image rotations. It is experimentally shown that this boost is obtained without reducing performance on ordinary illumination and viewpoint matching sequences.

Key facts. Cited by 27 papers in the map (data through 7 Oct 2026). Builds on 33 of its 57 references that are in the map. First posted on arXiv 21 Apr 2022.

Builds on

33 of 57 references.

References · 33

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2016Feature Pyramid Networks for Object Detectioncited by 5,021
  3. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,437
  4. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
  5. 2016Group Equivariant Convolutional Networkscited by 1,154
  6. 2021Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gaugescited by 1,131
  7. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  8. 2018MegaDepth: Learning Single-View Depth Prediction from Internet Photoscited by 731
  9. 2019General E(2)-Equivariant Steerable CNNscited by 441
  10. 20183D Steerable CNNs: Learning Rotationally Equivariant Features in Volumetric Datacited by 348
  11. 2019D2-Net: A Trainable CNN for Joint Detection and Description of Local Featurescited by 324
  12. 2020DISK: Learning local features with policy gradientcited by 310
21 more
  1. 2016LIFT: Learned Invariant Feature Transformcited by 305
  2. 2016Steerable CNNscited by 298
  3. 2018On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groupscited by 296
  4. 2019Gauge Equivariant Convolutional Networks and the Icosahedral CNNcited by 262
  5. 2017HPatches: A benchmark and evaluation of handcrafted and learned local descriptorscited by 261
  6. 2019R2D2: Repeatable and Reliable Detector and Descriptorcited by 206
  7. 2017Learning to Find Good Correspondencescited by 176
  8. 2019Learning Two-View Correspondences and Geometry Using Order-Aware Networkcited by 148
  9. 2020Patch2Pix: Epipolar-Guided Pixel-Level Correspondencescited by 128
  10. 2018Roto-Translation Covariant Convolutional Networks for Medical Image Analysiscited by 91
  11. 2020Dual-Resolution Correspondence Networkscited by 90
  12. 2021Coordinate Independent Convolutional Networks -- Isometry and Gauge Equivariant Convolutions on Riemannian Manifoldscited by 72
  13. 2021Geometric Deep Learning and Equivariant Neural Networkscited by 64
  14. 2019GIFT: Learning Transformation-Invariant Dense Visual Descriptors via Group CNNscited by 47
  15. 2020A Wigner-Eckart Theorem for Group Equivariant Convolution Kernelscited by 46
  16. 2021Warp Consistency for Unsupervised Learning of Dense Correspondencescited by 42
  17. 2016Learning Covariant Feature Detectorscited by 25
  18. 2020Single-Image Depth Prediction Makes Feature Matching Easiercited by 21
  19. 2021Homogeneous vector bundles and G-equivariant convolutional neural networkscited by 21
  20. 2015Learning to Assign Orientations to Feature Pointscited by 16
  21. 2021ZZ-Net: A Universal Rotation Equivariant Architecture for 2D Point Cloudscited by 15

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 4; 2023: 7; 2024: 6; 2025: 6; 2026: 4.

27 papers.

Sort:

  1. 2024Grounding Image Matching in 3D with MASt3Rcited by 960
  2. 2023RoMa: Robust Dense Feature Matchingcited by 296
  3. 2023DeDoDe: Detect, Don't Describe -- Describe, Don't Detect for Local Feature Matchingcited by 62
  4. 2024DeDoDe v2: Analyzing and Improving the DeDoDe Keypoint Detectorcited by 38
  5. 2024Local Feature Matching Using Deep Learning: A Surveycited by 26
  6. 2023Steerers: A framework for rotation equivariant keypoint descriptorscited by 21
  7. 2024EarthMatch: Iterative Coregistration for Fine-grained Localization of Astronaut Photographycited by 13
  8. 2023Learning Structure-from-Motion with Graph Attention Networkscited by 10
  9. 2023Searching from Area to Point: A Hierarchical Framework for Semantic-Geometric Combined Feature Matchingcited by 9
  10. 2023Investigating how ReLU-networks encode symmetriescited by 8
  11. 2024Affine steerers for structured keypoint descriptioncited by 6
  12. 2023Image Matching by Bare Homographycited by 5
15 more
  1. 2024Image Matching Filtering and Refinement by Planes and Beyondcited by 5
  2. 2022DRKF: Distilled Rotated Kernel Fusion for Efficient Rotation Invariant Descriptors in Local Feature Matchingcited by 3
  3. 2025To Match or Not to Match: Revisiting Image Matching for Reliable Visual Place Recognitioncited by 3
  4. 2025Quick ViTs: Speeding up Vision Transformers through Equivariancecited by 3
  5. 2022Descriptor Distillation: a Teacher-Student-Regularized Framework for Learning Local Descriptorscited by 2
  6. 2025ColabSfM: Collaborative Structure-from-Motion by Point Cloud Registrationcited by 2
  7. 2026Who Handles Orientation? Investigating Invariance in Feature Matchingcited by 1
  8. 2022SIM2E: Benchmarking the Group Equivariant Capability of Correspondence Matching Algorithmscited by 0
  9. 2022In Search of Projectively Equivariant Networkscited by 0
  10. 2025ISS-Geo142: A Benchmark for Geolocating Astronaut Photography from the International Space Stationcited by 0
  11. 2025Moment kernels: a simple and scalable approach for equivariance to rotations and reflections in deep convolutional networkscited by 0
  12. 2025EndoMatcher: Generalizable Endoscopic Image Matcher via Multi-Domain Pre-training for Robot-Assisted Surgerycited by 0
  13. 2026AsymLoc: Towards Asymmetric Feature Matching for Efficient Visual Localizationcited by 0
  14. 2026REDI-Match: Rotation-Equivariant Distillation for Efficient and Robust Dense Matchingcited by 0
  15. 2026A Vision Based System for Guided and Collaborative Reconstruction of Fragmented Documentscited by 0