{"url":"/task/scene-flow-estimation","name":"Scene Flow Estimation","slug":"scene-flow-estimation","description_markdown":"Optical flow is a two-dimensional motion field in the image plane. It is the projection of the three-dimensional motion of the world. If the world is completely non-rigid, the motions of the points in the scene may all be indepen- dent of each other. One representation of the scene motion is therefore a dense **three-dimensional vector field** defined for every point on every surface in the scene. By analogy with optical flow, we refer to this three-dimensional motion field as **scene flow**.\r\n\r\nSource: Vedula, Sundar, et al. \"Three-dimensional scene flow.\" IEEE transactions on pattern analysis and machine intelligence 27.3 (2005): 475-480. [pdf](https://ieeexplore.ieee.org/document/1388274)","categories":[{"name":"Computer Vision","url":"/area/computer-vision"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":152,"papers_with_code":81,"benchmarks":5,"benchmark_tables_in_archive":5,"benchmark_tables_shown":5,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":8,"subtasks":1,"parent_tasks":0},"benchmarks":[{"leaderboard":"/sota/scene-flow-estimation-on-argoverse-2","slug":"scene-flow-estimation-on-argoverse-2","dataset":"Argoverse 2","dataset_url":"/dataset/argoverse-2","rows_in_archive":7,"metrics":["EPE 3-Way","EPE Foreground Dynamic","EPE Foreground Static","EPE Background Static"],"first_row_in_archive_order":{"model":"DeFlow","paper_title":"DeFlow: Decoder of Scene Flow Network in Autonomous Driving","paper_url":"/paper/deflow-decoder-of-scene-flow-network-in","paper_date":"2024-01-29","arxiv_id":"2401.16122","code_links":[{"title":"kth-rpl/dynamicmap_benchmark","url":"https://github.com/kth-rpl/dynamicmap_benchmark"},{"title":"kth-rpl/deflow","url":"https://github.com/kth-rpl/deflow"},{"title":"kth-rpl/seflow","url":"https://github.com/kth-rpl/seflow"},{"title":"tudelft-iv/voteflow","url":"https://github.com/tudelft-iv/voteflow"}],"syntology":{"n":5,"n_ran":0,"n_unverified":5,"n_pointer_only":0}}},{"leaderboard":"/sota/scene-flow-estimation-on-spring","slug":"scene-flow-estimation-on-spring","dataset":"Spring","dataset_url":"/dataset/spring","rows_in_archive":6,"metrics":["1px total"],"first_row_in_archive_order":{"model":"M-FUSE (F)","paper_title":"M-FUSE: Multi-frame Fusion for Scene Flow Estimation","paper_url":"/paper/m-fuse-multi-frame-fusion-for-scene-flow","paper_date":"2022-07-12","arxiv_id":"2207.05704","code_links":[{"title":"cv-stuttgart/m-fuse","url":"https://github.com/cv-stuttgart/m-fuse"}],"syntology":null}},{"leaderboard":"/sota/scene-flow-estimation-on-kitti-2015-scene","slug":"scene-flow-estimation-on-kitti-2015-scene","dataset":"KITTI 2015 Scene Flow Training","dataset_url":"/dataset/kitti","rows_in_archive":4,"metrics":["D1-all","D2-all","Fl-all","SF-all"," Runtime (s)"],"first_row_in_archive_order":{"model":"EPC++","paper_title":"Every Pixel Counts ++: Joint Learning of Geometry and Motion with 3D Holistic Understanding","paper_url":"/paper/every-pixel-counts-joint-learning-of-geometry","paper_date":"2018-10-14","arxiv_id":"1810.06125","code_links":[{"title":"chenxuluo/EPC","url":"https://github.com/chenxuluo/EPC"}],"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":0}}},{"leaderboard":"/sota/scene-flow-estimation-on-kitti-2015-scene-1","slug":"scene-flow-estimation-on-kitti-2015-scene-1","dataset":"KITTI 2015 Scene Flow Test","dataset_url":"/dataset/kitti","rows_in_archive":4,"metrics":["SF-all","Fl-all","D1-all","D2-all","Runtime (s)"],"first_row_in_archive_order":{"model":"CamLiRAFT","paper_title":"Learning Optical Flow and Scene Flow with Bidirectional Camera-LiDAR Fusion","paper_url":"/paper/learning-optical-flow-and-scene-flow-with","paper_date":"2023-03-21","arxiv_id":"2303.12017","code_links":[{"title":"mcg-nju/camliflow","url":"https://github.com/mcg-nju/camliflow"}],"syntology":null}},{"leaderboard":"/sota/scene-flow-estimation-on-scene-flow","slug":"scene-flow-estimation-on-scene-flow","dataset":"Scene Flow","dataset_url":null,"rows_in_archive":1,"metrics":["EPE"],"first_row_in_archive_order":{"model":"AANet","paper_title":"AANet: Adaptive Aggregation Network for Efficient Stereo Matching","paper_url":"/paper/aanet-adaptive-aggregation-network-for","paper_date":"2020-04-20","arxiv_id":"2004.09548","code_links":[{"title":"haofeixu/aanet","url":"https://github.com/haofeixu/aanet"}],"syntology":{"n":14,"n_ran":9,"n_unverified":5,"n_pointer_only":0}}}],"datasets":[{"url":"/dataset/kitti","name":"KITTI","full_name":"","num_papers_in_archive":3661},{"url":"/dataset/flyingthings3d","name":"FlyingThings3D","full_name":"","num_papers_in_archive":226},{"url":"/dataset/argoverse-2","name":"Argoverse 2","full_name":"","num_papers_in_archive":185},{"url":"/dataset/deformingthings4d","name":"DeformingThings4D","full_name":"","num_papers_in_archive":42},{"url":"/dataset/spring","name":"Spring","full_name":"Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo","num_papers_in_archive":29},{"url":"/dataset/vbr","name":"VBR","full_name":"VBR: A Vision Benchmark in Rome","num_papers_in_archive":3},{"url":"/dataset/middlebury-2003","name":"Middlebury 2003","full_name":"Middlebury 2003","num_papers_in_archive":2},{"url":"/dataset/ekubric","name":"EKubric","full_name":"","num_papers_in_archive":1}],"subtasks":[{"url":"/task/self-supervised-scene-flow-estimation","name":"Self-supervised Scene Flow Estimation"}],"parent_tasks":[],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":81,"tagged_in_all":152,"items":[{"url":"/paper/scalable-scene-flow-from-point-clouds-in-the","title":"Scalable Scene Flow from Point Clouds in the Real World","date":"2021-03-01","arxiv_id":"2103.01306","repositories_listed":5,"syntology":null},{"url":"/paper/deflow-decoder-of-scene-flow-network-in","title":"DeFlow: Decoder of Scene Flow Network in Autonomous Driving","date":"2024-01-29","arxiv_id":"2401.16122","repositories_listed":4,"syntology":{"n":5,"n_ran":0,"n_unverified":5,"n_pointer_only":0}},{"url":"/paper/a-large-dataset-to-train-convolutional","title":"A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow Estimation","date":"2015-12-07","arxiv_id":"1512.02134","repositories_listed":3,"syntology":null},{"url":"/paper/seflow-a-self-supervised-scene-flow-method-in","title":"SeFlow: A Self-Supervised Scene Flow Method in Autonomous Driving","date":"2024-07-01","arxiv_id":"2407.01702","repositories_listed":2,"syntology":{"n":2,"n_ran":0,"n_unverified":2,"n_pointer_only":0}},{"url":"/paper/let-it-flow-simultaneous-optimization-of-3d","title":"Let-It-Flow: Simultaneous Optimization of 3D Flow and Object Clustering","date":"2024-04-12","arxiv_id":"2404.08363","repositories_listed":2,"syntology":null},{"url":"/paper/milliflow-scene-flow-estimation-on-mmwave","title":"milliFlow: Scene Flow Estimation on mmWave Radar Point Cloud for Human Motion Sensing","date":"2023-06-29","arxiv_id":"2306.17010","repositories_listed":2,"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":3}},{"url":"/paper/spring-a-high-resolution-high-detail-dataset","title":"Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo","date":"2023-03-03","arxiv_id":"2303.01943","repositories_listed":2,"syntology":{"n":6,"n_ran":4,"n_unverified":2,"n_pointer_only":4}},{"url":"/paper/3d-object-detection-with-a-self-supervised","title":"3D Object Detection with a Self-supervised Lidar Scene Flow Backbone","date":"2022-05-02","arxiv_id":"2205.00705","repositories_listed":2,"syntology":{"n":16,"n_ran":0,"n_unverified":16,"n_pointer_only":0}},{"url":"/paper/self-supervised-scene-flow-estimation-with-4d","title":"Self-Supervised Scene Flow Estimation with 4-D Automotive Radar","date":"2022-03-02","arxiv_id":"2203.01137","repositories_listed":2,"syntology":{"n":12,"n_ran":2,"n_unverified":10,"n_pointer_only":0}},{"url":"/paper/accurate-point-cloud-registration-with-robust","title":"Accurate Point Cloud Registration with Robust Optimal Transport","date":"2021-11-01","arxiv_id":"2111.00648","repositories_listed":2,"syntology":null},{"url":"/paper/pointpwc-net-a-coarse-to-fine-network-for","title":"PointPWC-Net: A Coarse-to-Fine Network for Supervised and Self-Supervised Scene Flow Estimation on 3D Point Clouds","date":"2019-11-27","arxiv_id":"1911.12408","repositories_listed":2,"syntology":{"n":18,"n_ran":2,"n_unverified":16,"n_pointer_only":0}},{"url":"/paper/meteornet-deep-learning-on-dynamic-3d-point","title":"MeteorNet: Deep Learning on Dynamic 3D Point Cloud Sequences","date":"2019-10-21","arxiv_id":"1910.09165","repositories_listed":2,"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/hplflownet-hierarchical-permutohedral-lattice-1","title":"HPLFlowNet: Hierarchical Permutohedral Lattice FlowNet for Scene Flow Estimation on Large-scale Point Clouds","date":"2019-06-12","arxiv_id":"1906.05332","repositories_listed":2,"syntology":null},{"url":"/paper/kdmos-knowledge-distillation-for-motion","title":"KDMOS:Knowledge Distillation for Motion Segmentation","date":"2025-06-17","arxiv_id":"2506.14130","repositories_listed":1,"syntology":null},{"url":"/paper/voteflow-enforcing-local-rigidity-in-self","title":"VoteFlow: Enforcing Local Rigidity in Self-Supervised Scene Flow","date":"2025-03-28","arxiv_id":"2503.22328","repositories_listed":1,"syntology":null},{"url":"/paper/mambaflow-a-novel-and-flow-guided-state-space","title":"MambaFlow: A Novel and Flow-guided State Space Model for Scene Flow Estimation","date":"2025-02-24","arxiv_id":"2502.16907","repositories_listed":1,"syntology":{"n":1,"n_ran":0,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/ssf-sparse-long-range-scene-flow-for","title":"SSF: Sparse Long-Range Scene Flow for Autonomous Driving","date":"2025-01-29","arxiv_id":"2501.17821","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/storm-spatio-temporal-reconstruction-model","title":"STORM: Spatio-Temporal Reconstruction Model for Large-Scale Outdoor Scenes","date":"2024-12-31","arxiv_id":"2501.00602","repositories_listed":1,"syntology":null},{"url":"/paper/fastpci-motion-structure-guided-fast-point","title":"FastPCI: Motion-Structure Guided Fast Point Cloud Frame Interpolation","date":"2024-10-25","arxiv_id":"2410.19573","repositories_listed":1,"syntology":null},{"url":"/paper/scaleflow-robust-and-accurate-estimation-of","title":"ScaleFlow++: Robust and Accurate Estimation of 3D Motion from Video","date":"2024-09-16","arxiv_id":"2409.12202","repositories_listed":1,"syntology":null},{"url":"/paper/scaleraft-cross-scale-recurrent-all-pairs","title":"ScaleFlow++: Robust and Accurate Estimation of 3D Motion from Video","date":"2024-07-13","arxiv_id":"2407.09797","repositories_listed":1,"syntology":null},{"url":"/paper/flow4d-leveraging-4d-voxel-network-for-lidar","title":"Flow4D: Leveraging 4D Voxel Network for LiDAR Scene Flow Estimation","date":"2024-07-10","arxiv_id":"2407.07995","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/scenetracker-long-term-scene-flow-estimation","title":"SceneTracker: Long-term Scene Flow Estimation Network","date":"2024-03-29","arxiv_id":"2403.19924","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":0}},{"url":"/paper/fast-kernel-scene-flow","title":"Fast Kernel Scene Flow","date":"2024-03-09","arxiv_id":"2403.05896","repositories_listed":1,"syntology":null},{"url":"/paper/diffsf-diffusion-models-for-scene-flow","title":"DiffSF: Diffusion Models for Scene Flow Estimation","date":"2024-03-08","arxiv_id":"2403.05327","repositories_listed":1,"syntology":{"n":11,"n_ran":10,"n_unverified":1,"n_pointer_only":11}},{"url":"/paper/3dsflabelling-boosting-3d-scene-flow","title":"3DSFLabelling: Boosting 3D Scene Flow Estimation by Pseudo Auto-labelling","date":"2024-02-28","arxiv_id":"2402.18146","repositories_listed":1,"syntology":{"n":10,"n_ran":9,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/icp-flow-lidar-scene-flow-estimation-with-icp","title":"ICP-Flow: LiDAR Scene Flow Estimation with ICP","date":"2024-02-27","arxiv_id":"2402.17351","repositories_listed":1,"syntology":{"n":9,"n_ran":8,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/fedrsu-federated-learning-for-scene-flow","title":"FedRSU: Federated Learning for Scene Flow Estimation on Roadside Units","date":"2024-01-23","arxiv_id":"2401.12862","repositories_listed":1,"syntology":null},{"url":"/paper/difflow3d-toward-robust-uncertainty-aware-1","title":"DifFlow3D: Toward Robust Uncertainty-Aware Scene Flow Estimation with Iterative Diffusion-Based Refinement","date":"2024-01-01","arxiv_id":null,"repositories_listed":1,"syntology":null},{"url":"/paper/regularizing-self-supervised-3d-scene-flows","title":"Regularizing Self-supervised 3D Scene Flows with Surface Awareness and Cyclic Consistency","date":"2023-12-12","arxiv_id":"2312.08879","repositories_listed":1,"syntology":null}],"syntology_records":15,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}