mlsanityData to 7 Oct 2026

Warp Consistency for Unsupervised Learning of Dense Correspondences

arXiv 2104.03308 ↗First posted 7 Apr 2021cs.CVcited by 42

Not on the map. Why?

Abstract (arXiv)

The key challenge in learning dense correspondences lies in the lack of ground-truth matches for real image pairs. While photometric consistency losses provide unsupervised alternatives, they struggle with large appearance changes, which are ubiquitous in geometric and semantic matching tasks. Moreover, methods relying on synthetic training pairs often suffer from poor generalisation to real data. We propose Warp Consistency, an unsupervised learning objective for dense correspondence regression. Our objective is effective even in settings with large appearance and view-point changes. Given a pair of real images, we first construct an image triplet by applying a randomly sampled warp to one of the original images. We derive and analyze all flow-consistency constraints arising between the triplet. From our observations and empirical results, we design a general unsupervised objective employing two of the derived constraints. We validate our warp consistency loss by training three recent dense correspondence networks for the geometric and semantic matching tasks. Our approach sets a new state-of-the-art on several challenging benchmarks, including MegaDepth, RobotCar and TSS. Code and models are at github.com/PruneTruong/DenseMatching.

Key facts. Cited by 42 papers in the map (data through 7 Oct 2026). Builds on 55 of its 85 references that are in the map. First posted on arXiv 7 Apr 2021.

Builds on

55 of 85 references.

References · 55

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  4. 2016Densely Connected Convolutional Networkscited by 6,691
  5. 2016The Cityscapes Dataset for Semantic Urban Scene Understandingcited by 4,427
  6. 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,313
  7. 2014Two-Stream Convolutional Networks for Action Recognition in Videoscited by 1,067
  8. 2016Semantic Understanding of Scenes through the ADE20K Datasetcited by 1,063
  9. 2015FlowNet: Learning Optical Flow with Convolutional Networkscited by 1,054
  10. 2017PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volumecited by 966
  11. 2015A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimationcited by 878
  12. 2016FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networkscited by 780
38 more
  1. 2018MegaDepth: Learning Single-View Depth Prediction from Internet Photoscited by 731
  2. 2018Albumentations: fast and flexible image augmentationscited by 558
  3. 2019D2-Net: A Trainable CNN for Joint Detection and Description of Local Featurescited by 324
  4. 2019Learning Correspondence from the Cycle-Consistency of Timecited by 269
  5. 2017HPatches: A benchmark and evaluation of handcrafted and learned local descriptorscited by 261
  6. 2020Space-Time Correspondence as a Contrastive Random Walkcited by 239
  7. 2017UnFlow: Unsupervised Learning of Optical Flow with a Bidirectional Census Losscited by 202
  8. 2018Neighbourhood Consensus Networkscited by 173
  9. 2017DSLR-Quality Photos on Mobile Devices with Deep Convolutional Networkscited by 156
  10. 2019Temporal Cycle-Consistency Learningcited by 151
  11. 2019Learning Two-View Correspondences and Geometry Using Order-Aware Networkcited by 148
  12. 2019SPair-71k: A Large-scale Benchmark for Semantic Correspondencecited by 140
  13. 2019GLU-Net: Global-Local Universal Network for Dense Flow and Correspondencescited by 139
  14. 2017Convolutional neural network architecture for geometric matchingcited by 138
  15. 2019SelFlow: Self-Supervised Learning of Optical Flowcited by 128
  16. 2017Visual Attribute Transfer through Deep Image Analogycited by 123
  17. 2014Return of the Devil in the Details: Delving Deep into Convolutional Netscited by 122
  18. 2020Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimationcited by 120
  19. 2016Back to Basics: Unsupervised Learning of Optical Flow via Brightness Constancy and Motion Smoothnesscited by 101
  20. 2017Occlusion Aware Unsupervised Learning of Optical Flowcited by 101
  21. 2020Efficient Neighbourhood Consensus Networks via Submanifold Sparse Convolutionscited by 92
  22. 2016Learning Dense Correspondence via 3D-guided Cycle Consistencycited by 91
  23. 2020Dual-Resolution Correspondence Networkscited by 90
  24. 2018DGC-Net: Dense Geometric Correspondence Networkcited by 67
  25. 2019DDFlow: Learning Optical Flow with Unlabeled Data Distillationcited by 63
  26. 2017End-to-end weakly-supervised semantic alignmentcited by 60
  27. 2019SFNet: Learning Object-aware Semantic Correspondencecited by 60
  28. 2015Proposal Flowcited by 57
  29. 2020Learning to Compose Hypercolumns for Visual Correspondencecited by 53
  30. 2020GOCor: Bringing Globally Optimized Correspondence Volumes into Your Neural Networkcited by 53
  31. 2017SCNet: Learning Semantic Correspondencecited by 48
  32. 2020RANSAC-Flow: generic two-stage image alignmentcited by 48
  33. 2017Proposal Flow: Semantic Correspondences from Object Proposalscited by 46
  34. 2019Dynamic Context Correspondence Network for Semantic Alignmentcited by 43
  35. 2018Recurrent Transformer Networks for Semantic Correspondencecited by 32
  36. 2018Attentive Semantic Alignment with Offset-Aware Correlation Kernelscited by 31
  37. 2019GLAMpoints: Greedily Learned Accurate Match pointscited by 30
  38. 2019A Cross-Season Correspondence Dataset for Robust Semantic Segmentationcited by 28

50 most cited of 55 All 55 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 8; 2023: 12; 2024: 8; 2025: 9; 2026: 4.

42 papers.

Sort:

By area:

  1. 2023Emergent Correspondence from Image Diffusioncited by 437
  2. 2023A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondencecited by 297
  3. 2023Telling Left from Right: Identifying Geometry-Aware Semantic Correspondencecited by 85
  4. 2024DreamMatcher: Appearance Matching Self-Attention for Semantically-Consistent Text-to-Image Personalizationcited by 60
  5. 2024DeDoDe v2: Analyzing and Improving the DeDoDe Keypoint Detectorcited by 38
  6. 2022Learning Pixel Trajectories with Multiscale Contrastive Random Walkscited by 36
  7. 2022Refign: Align and Refine for Adaptation of Semantic Segmentation to Adverse Conditionscited by 31
  8. 2022One-Shot Transfer of Affordance Regions? AffCorrs!cited by 31
  9. 2022Probabilistic Warp Consistency for Weakly-Supervised Semantic Correspondencescited by 30
  10. 2022A case for using rotation invariant features in state of the art feature matcherscited by 27
  11. 2023ASIC: Aligning Sparse in-the-wild Image Collectionscited by 26
  12. 2022Self-Supervised Geometric Correspondence for Category-Level 6D Object Pose Estimation in the Wildcited by 22
30 more
  1. 2023Chain of Visual Perception: Harnessing Multimodal Large Language Models for Zero-shot Camouflaged Object Detectioncited by 22
  2. 2021Learning Co-segmentation by Segment Swapping for Retrieval and Discoverycited by 15
  3. 2023Towards Training-free Open-world Segmentation via Image Prompt Foundation Modelscited by 15
  4. 2025Do It Yourself: Learning Semantic Correspondence from Pseudo-Labelscited by 13
  5. 2024SHIC: Shape-Image Correspondences with no Keypoint Supervisioncited by 11
  6. 2022Semi-supervised Deep Large-baseline Homography Estimation with Progressive Equivalence Constraintcited by 10
  7. 2023GyroFlow+: Gyroscope-Guided Unsupervised Deep Homography and Optical Flow Learningcited by 10
  8. 2024Zero-Shot Image Feature Consensus with Deep Functional Mapscited by 10
  9. 2025Consistent Image Layout Editing with Diffusion Modelscited by 9
  10. 2025Deep Learning Reforms Image Matching: A Survey and Outlookcited by 7
  11. 2023Doduo: Learning Dense Visual Correspondence from Unsupervised Semantic-Aware Flowcited by 6
  12. 2023Warping the Residuals for Image Editing with StyleGANcited by 6
  13. 2023Towards Content-based Pixel Retrieval in Revisited Oxford and Pariscited by 5
  14. 2025Jamais Vu: Exposing the Generalization Gap in Supervised Semantic Correspondencecited by 5
  15. 2023Zero-Shot Co-salient Object Detection Frameworkcited by 4
  16. 2024CONDA: Condensed Deep Association Learning for Co-Salient Object Detectioncited by 4
  17. 2025SketchFusion: Learning Universal Sketch Features through Fusing Foundation Modelscited by 4
  18. 2025Semantic Correspondence: Unified Benchmarking and a Strong Baselinecited by 3
  19. 2025Towards Geometry-Grounded Dense Semantic Matching with VGGT Priorscited by 3
  20. 2023Match me if you can: Semi-Supervised Semantic Correspondence Learning with Unpaired Imagescited by 2
  21. 2025Self-Supervised Spatial Correspondence Across Modalitiescited by 2
  22. 2026Personalized Image Generation via Human-in-the-loop Bayesian Optimizationcited by 2
  23. 2022Self-supervised Pre-training for Semantic Segmentation in an Indoor Scenecited by 1
  24. 2024CARL: A Framework for Equivariant Image Registrationcited by 1
  25. 2024Cycle-Correspondence Loss: Learning Dense View-Invariant Visual Features from Unlabeled and Unordered RGB Imagescited by 1
  26. 2026CORAL: Correspondence Alignment for Improved Virtual Try-Oncited by 1
  27. 2024Accurate and Fast Pixel Retrieval with Spatial and Uncertainty Aware Hypergraph Diffusioncited by 0
  28. 2025Blending 3D Geometry and Machine Learning for Multi-View Stereopsiscited by 0
  29. 2026Deep learning enables urban change profiling through alignment of historical mapscited by 0
  30. 2026Gromov Wasserstein Optimal Transport for Semantic Correspondencescited by 0