mlsanityData to 7 Oct 2026

SEA-RAFT: Simple, Efficient, Accurate RAFT for Optical Flow

arXiv 2405.14793 ↗First posted 23 May 2024cs.CVcited by 169

Not on the map. Why?

Abstract (arXiv)

We introduce SEA-RAFT, a more simple, efficient, and accurate RAFT for optical flow. Compared with RAFT, SEA-RAFT is trained with a new loss (mixture of Laplace). It directly regresses an initial flow for faster convergence in iterative refinements and introduces rigid-motion pre-training to improve generalization. SEA-RAFT achieves state-of-the-art accuracy on the Spring benchmark with a 3.69 endpoint-error (EPE) and a 0.36 1-pixel outlier rate (1px), representing 22.9% and 17.8% error reduction from best published results. In addition, SEA-RAFT obtains the best cross-dataset generalization on KITTI and Spring. With its high efficiency, SEA-RAFT operates at least 2.3x faster than existing methods while maintaining competitive performance. The code is publicly available at https://github.com/princeton-vl/SEA-RAFT.

Key facts. Cited by 169 papers in the map (data through 7 Oct 2026). Builds on 49 of its 70 references that are in the map. First posted on arXiv 23 May 2024.

Builds on

49 of 70 references.

References · 49

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 2022A ConvNet for the 2020scited by 3,545
  3. 2020RAFT: Recurrent All-Pairs Field Transforms for Optical Flowcited by 2,385
  4. 2015FlowNet: Learning Optical Flow with Convolutional Networkscited by 1,054
  5. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 974
  6. 2017PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volumecited by 966
  7. 2015A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimationcited by 878
  8. 2016FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networkscited by 780
  9. 2018MegaDepth: Learning Single-View Depth Prediction from Internet Photoscited by 731
  10. 2020TartanAir: A Dataset to Push the Limits of Visual SLAMcited by 494
  11. 2018YouTube-VOS: A Large-Scale Video Object Segmentation Benchmarkcited by 381
  12. 2021GMFlow: Learning Optical Flow via Global Matchingcited by 324
37 more
  1. 2022FlowFormer: A Transformer Architecture for Optical Flowcited by 273
  2. 2022Unifying Flow, Stereo and Depth Estimationcited by 270
  3. 2021Learning to Estimate Hidden Motions with Global Motion Aggregationcited by 223
  4. 2022ASpanFormer: Detector-Free Image Matching with Adaptive Span Transformercited by 170
  5. 2023Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereocited by 170
  6. 2017Playing for Benchmarkscited by 169
  7. 2022CroCo v2: Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flowcited by 165
  8. 2022CroCo: Self-Supervised Pre-training for 3D Vision Tasks by Cross-View Completioncited by 158
  9. 2019Quadratic video interpolationcited by 143
  10. 2023Infinite Photorealistic Worlds using Procedural Generationcited by 142
  11. 2021Human Pose Regression with Residual Log-likelihood Estimationcited by 131
  12. 2020MaskFlownet: Asymmetric Feature Matching with Learnable Occlusion Maskcited by 117
  13. 2019Deep Flow-Guided Video Inpaintingcited by 116
  14. 2023The Surprising Effectiveness of Diffusion Models for Optical Flow and Monocular Depth Estimationcited by 116
  15. 2023FlowFormer++: Masked Cost Volume Autoencoding for Pretraining Optical Flow Estimationcited by 114
  16. 2019Deep Video Inpaintingcited by 107
  17. 2020Flow-edge Guided Video Completioncited by 104
  18. 2018Models Matter, So Does Training: An Empirical Study of CNNs for Optical Flow Estimationcited by 94
  19. 2017Optical Flow Guided Feature: A Fast and Robust Motion Representation for Video Action Recognitioncited by 82
  20. 2022CRAFT: Cross-Attentional Flow Transformer for Robust Optical Flowcited by 82
  21. 2021PDC-Net+: Enhanced Probabilistic Dense Correspondence Networkcited by 80
  22. 2017Accurate Optical Flow via Direct Cost Volume Processingcited by 66
  23. 2022SKFlow: Learning Optical Flow with Super Kernelscited by 56
  24. 2018Representation Flow for Action Recognitioncited by 51
  25. 2022Multiview Stereo with Cascaded Epipolar RAFTcited by 51
  26. 2022DIP: Deep Inverse Patchmatch for High-Resolution Optical Flowcited by 40
  27. 2016Full Flow: Optical Flow Estimation By Global Optimization over Regular Gridscited by 34
  28. 2022Learning Optical Flow with Adaptive Graph Reasoningcited by 33
  29. 2023Rethinking Optical Flow from Geometric Matching Consistent Perspectivecited by 33
  30. 2023TransFlow: Transformer as Flow Learnercited by 32
  31. 2017ProbFlow: Joint Optical Flow and Uncertainty Estimationcited by 24
  32. 2022View Synthesis with Sculpted Neural Pointscited by 20
  33. 2023GAFlow: Incorporating Gaussian Attention into Optical Flowcited by 18
  34. 2023AnyFlow: Arbitrary Scale Optical Flow with Implicit Neural Representationcited by 17
  35. 2023CCMR: High Resolution Optical Flow Estimation via Coarse-to-Fine Context-Guided Motion Reasoningcited by 15
  36. 2023Win-Win: Training High-Resolution Vision Transformers from Two Windowscited by 12
  37. 2023DIFT: Dynamic Iterative Field Transforms for Memory Efficient Optical Flowcited by 7

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 13; 2025: 74; 2026: 82.

169 papers.

Sort:

By area:

  1. 2024MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motioncited by 431
  2. 2025WorldScore: A Unified Evaluation Benchmark for World Generationcited by 143
  3. 2024Stereo4D: Learning How Things Move in 3D from Internet Stereo Videoscited by 89
  4. 2025Easi3R: Estimating Disentangled Motion from DUSt3R Without Trainingcited by 88
  5. 2025Towards Practical Real-Time Neural Video Compressioncited by 72
  6. 2024DELTA: Dense Efficient Long-range 3D Tracking for any videocited by 62
  7. 2025AllTracker: Efficient Dense Point Tracking at High Resolutioncited by 54
  8. 2025WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environmentscited by 48
  9. 2025UFM: A Simple Path towards Unified Dense Correspondence with Flowcited by 35
  10. 2025WAFT: Warping-Alone Field Transforms for Optical Flowcited by 32
  11. 2025Any4D: Unified Feed-Forward Metric 4D Reconstructioncited by 32
  12. 2025Trace Anything: Representing Any Video in 4D via Trajectory Fieldscited by 31
38 more
  1. 2024Cross-View Completion Models are Zero-shot Correspondence Estimatorscited by 30
  2. 2025DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Frameworkcited by 30
  3. 2024OMNI-DC: Highly Robust Depth Completion with Multiresolution Depth Integrationcited by 28
  4. 2025POMATO: Marrying Pointmap Matching with Temporal Motion for Dynamic 3D Reconstructioncited by 24
  5. 2026WorldMark: A Unified Benchmark Suite for Interactive Video World Modelscited by 24
  6. 2025VideoGen-Eval: Agent-based System for Video Generation Evaluationcited by 23
  7. 2025Advances in Feed-Forward 3D Reconstruction and View Synthesis: A Surveycited by 23
  8. 2024SplatFlow: Self-Supervised Dynamic Gaussian Splatting in Neural Motion Flow Field for Autonomous Drivingcited by 20
  9. 2026DrivingGen: A Comprehensive Benchmark for Generative Video World Models in Autonomous Drivingcited by 20
  10. 2026Motion 3-to-4: 3D Motion Reconstruction for 4D Synthesiscited by 15
  11. 2026Vista4D: Video Reshooting with 4D Point Cloudscited by 15
  12. 2024Dynamics-Aware Gaussian Splatting Streaming Towards Fast On-the-Fly 4D Reconstructioncited by 14
  13. 2025WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perceptioncited by 14
  14. 2025FramePainter: Endowing Interactive Image Editing with Video Diffusion Priorscited by 13
  15. 20253D Scene Prompting for Scene-Consistent Camera-Controllable Video Generationcited by 13
  16. 2026CoWTracker: Tracking by Warping instead of Correlationcited by 12
  17. 2025Can Video Diffusion Model Reconstruct 4D Geometry?cited by 10
  18. 2025Scene-Centric Unsupervised Panoptic Segmentationcited by 10
  19. 2025D2USt3R: Enhancing 3D Reconstruction for Dynamic Scenescited by 10
  20. 2025FlowSeek: Optical Flow Made Easier with Depth Foundation Models and Motion Basescited by 10
  21. 2026Track4World: Feedforward World-centric Dense 3D Tracking of All Pixelscited by 10
  22. 2025Flow-Anything: Learning Real-World Optical Flow Estimation from Large-Scale Single-view Imagescited by 9
  23. 2025Flow4Agent: Long-form Video Understanding via Motion Prior from Optical Flowcited by 9
  24. 2025SMRABooth: Subject and Motion Representation Alignment for Customized Video Generationcited by 9
  25. 2024C-DiffSET: Leveraging Latent Diffusion for SAR-to-EO Image Translation with Confidence-Guided Reliable Object Generationcited by 8
  26. 2025Image Editing with Diffusion Models: A Surveycited by 8
  27. 2025A Unified Solution to Video Fusion: From Multi-Frame Learning to Benchmarkingcited by 8
  28. 2025Web-Scale Collection of Video Data for 4D Animal Reconstructioncited by 8
  29. 2026AdaGaR: Adaptive Gabor Representation for Dynamic Scene Reconstructioncited by 8
  30. 2026MLLM-4D: Towards Visual-based Spatial-Temporal Intelligencecited by 8
  31. 2026MoRe: Motion-aware Feed-forward 4D Reconstruction Transformercited by 8
  32. 2026Attend Before Attention: Efficient and Scalable Video Understanding via Autoregressive Gazingcited by 8
  33. 2026Generative World Renderercited by 8
  34. 2024MFTIQ: Multi-Flow Tracker with Independent Matching Quality Estimationcited by 7
  35. 2025Hierarchical Flow Diffusion for Efficient Frame Interpolationcited by 7
  36. 2025RobustSpring: Benchmarking Robustness to Image Corruptions for Optical Flow, Scene Flow and Stereocited by 7
  37. 2025World Modeling with Probabilistic Structure Integrationcited by 7
  38. 2025Forge4D: Feed-Forward 4D Human Reconstruction and Interpolation from Uncalibrated Sparse-view Videoscited by 7

50 most cited of 169 All 169 (JSON)