mlsanityData to 7 Oct 2026

Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimation

arXiv 2003.13045 ↗First posted 29 Mar 2020cs.CVcited by 120

Not on the map. Why?

Abstract (arXiv)

Unsupervised learning of optical flow, which leverages the supervision from view synthesis, has emerged as a promising alternative to supervised methods. However, the objective of unsupervised learning is likely to be unreliable in challenging scenes. In this work, we present a framework to use more reliable supervision from transformations. It simply twists the general unsupervised learning pipeline by running another forward pass with transformed data from augmentation, along with using transformed predictions of original data as the self-supervision signal. Besides, we further introduce a lightweight network with multiple frames by a highly-shared flow decoder. Our method consistently gets a leap of performance on several benchmarks with the best accuracy among deep unsupervised methods. Also, our method achieves competitive results to recent fully supervised methods while with much fewer parameters.

Key facts. Cited by 120 papers in the map (data through 7 Oct 2026). First posted on arXiv 29 Mar 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 27; 2022: 25; 2023: 28; 2024: 17; 2025: 15; 2026: 8.

120 papers.

Sort:

By area:

  1. 2021Self-supervised Video Object Segmentation by Motion Groupingcited by 147
  2. 2021Learn2Reg: comprehensive multi-task medical image registration challenge, dataset and evaluation in the era of deep learningcited by 125
  3. 2024CoTracker3: Simpler and Better Point Tracking by Pseudo-Labelling Real Videoscited by 122
  4. 2024BootsTAP: Bootstrapped Training for Tracking-Any-Pointcited by 115
  5. 2022TokenCut: Segmenting Objects in Images and Videos with Self-supervised Transformer and Normalized Cutcited by 111
  6. 2021Asymmetric Bilateral Motion Estimation for Video Frame Interpolationcited by 101
  7. 2021Repurposing GANs for One-shot Semantic Part Segmentationcited by 88
  8. 2021PDC-Net+: Enhanced Probabilistic Dense Correspondence Networkcited by 80
  9. 2021SMURF: Self-Teaching Multi-Frame Unsupervised RAFT with Full-Image Warpingcited by 74
  10. 2022Segmenting Moving Objects via an Object-Centric Layered Representationcited by 63
  11. 2022Learning Dense and Continuous Optical Flow from an Event Cameracited by 59
  12. 2024MemFlow: Optical Flow Estimation and Prediction with Memorycited by 55
38 more
  1. 2023MC-JEPA: A Joint-Embedding Predictive Architecture for Self-Supervised Learning of Motion and Content Featurescited by 53
  2. 2021Spatio-Temporal Recurrent Networks for Event-Based Optical Flow Estimationcited by 50
  3. 2021The Emergence of Objectness: Learning Zero-Shot Segmentation from Videoscited by 45
  4. 2021Warp Consistency for Unsupervised Learning of Dense Correspondencescited by 42
  5. 2022Guess What Moves: Unsupervised Video and Image Segmentation by Anticipating Motioncited by 42
  6. 2022Learning Pixel Trajectories with Multiscale Contrastive Random Walkscited by 36
  7. 2023Tracking and Mapping in Medical Computer Vision: A Reviewcited by 36
  8. 2022Disentangling Architecture and Training for Optical Flowcited by 32
  9. 2021Digging into Uncertainty in Self-supervised Multi-view Stereocited by 31
  10. 2021Optical Flow Estimation for Spiking Cameracited by 29
  11. 2021Unsupervised Object Learning via Common Fatecited by 29
  12. 2021PVStereo: Pyramid Voting Module for End-to-End Self-Supervised Stereo Matchingcited by 28
  13. 2021Learning optical flow from still imagescited by 28
  14. 2024Diff-Tracker: Text-to-Image Diffusion Models are Unsupervised Trackerscited by 24
  15. 2023BiFormer: Learning Bilateral Motion Estimation via Bilateral Transformer for 4K Video Frame Interpolationcited by 23
  16. 2022RealFlow: EM-based Realistic Optical Flow Dataset Generation from Videoscited by 22
  17. 2021Learning to Track Objects from Unlabeled Videoscited by 21
  18. 2021Improving the Perceptual Quality of 2D Animation Interpolationcited by 20
  19. 2022Unsupervised Multi-object Segmentation by Predicting Probable Motion Patternscited by 19
  20. 2023GAFlow: Incorporating Gaussian Attention into Optical Flowcited by 18
  21. 2022Unsupervised Learning of Accurate Siamese Trackingcited by 16
  22. 2021GyroFlow: Gyroscope-Guided Unsupervised Optical Flow Learningcited by 15
  23. 2022A Simple and Powerful Global Optimization for Unsupervised Video Object Segmentationcited by 15
  24. 2023Bootstrapping Objectness from Videos by Relaxed Common Fate and Visual Groupingcited by 15
  25. 2022Optical Flow Training under Limited Label Budget via Active Learningcited by 14
  26. 2023SurgT challenge: Benchmark of Soft-Tissue Trackers for Robotic Surgerycited by 14
  27. 2025Learning segmentation from point trajectoriescited by 14
  28. 2022Spatial-then-Temporal Self-Supervised Learning for Video Correspondencecited by 13
  29. 2021Occlusion Guided Self-supervised Scene Flow Estimation on 3D Point Cloudscited by 11
  30. 2022Unsupervised Flow-Aligned Sequence-to-Sequence Learning for Video Restorationcited by 11
  31. 2022Solve the Puzzle of Instance Segmentation in Videos: A Weakly Supervised Framework with Spatio-Temporal Collaborationcited by 11
  32. 2023Appearance-Based Refinement for Object-Centric Motion Segmentationcited by 11
  33. 2024Explicit Motion Handling and Interactive Prompting for Video Camouflaged Object Detectioncited by 11
  34. 2024Factorized Motion Fields for Fast Sparse Input Dynamic View Synthesiscited by 11
  35. 2022MDFlow: Unsupervised Optical Flow Learning by Reliable Mutual Knowledge Distillationcited by 10
  36. 2022Self-supervised AutoFlowcited by 10
  37. 2023GyroFlow+: Gyroscope-Guided Unsupervised Deep Homography and Optical Flow Learningcited by 10
  38. 2023Self-Supervised Motion Magnification by Backpropagating Through Optical Flowcited by 10

50 most cited of 120 All 120 (JSON)