SOTAVerified

Scene Flow Estimation

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.

Source: Vedula, Sundar, et al. "Three-dimensional scene flow." IEEE transactions on pattern analysis and machine intelligence 27.3 (2005): 475-480. pdf

Papers

Showing 150 of 152 papers

TitleStatusHype
DeFlow: Decoder of Scene Flow Network in Autonomous DrivingCode3
STORM: Spatio-Temporal Reconstruction Model for Large-Scale Outdoor ScenesCode3
SeFlow: A Self-Supervised Scene Flow Method in Autonomous DrivingCode2
Learning Optical Flow and Scene Flow with Bidirectional Camera-LiDAR FusionCode2
DifFlow3D: Toward Robust Uncertainty-Aware Scene Flow Estimation with Iterative Diffusion-Based RefinementCode2
SceneTracker: Long-term Scene Flow Estimation NetworkCode2
ICP-Flow: LiDAR Scene Flow Estimation with ICPCode2
SCOOP: Self-Supervised Correspondence and Optimization-Based Scene FlowCode1
Neighborhood Normalization for Robust Geometric Feature LearningCode1
SCTN: Sparse Convolution-Transformer Network for Scene Flow EstimationCode1
MonoPLFlowNet: Permutohedral Lattice FlowNet for Real-Scale 3D Scene FlowEstimation with Monocular ImagesCode1
milliFlow: Scene Flow Estimation on mmWave Radar Point Cloud for Human Motion SensingCode1
Accurate Point Cloud Registration with Robust Optimal TransportCode1
Multi-Body Neural Scene FlowCode1
ScaleFlow++: Robust and Accurate Estimation of 3D Motion from VideoCode1
Scalable Scene Flow from Point Clouds in the Real WorldCode1
ScaleFlow++: Robust and Accurate Estimation of 3D Motion from VideoCode1
PV-RAFT: Point-Voxel Correlation Fields for Scene Flow Estimation of Point CloudsCode1
Let-It-Flow: Simultaneous Optimization of 3D Flow and Object ClusteringCode1
LiteFlowNet3: Resolving Correspondence Ambiguity for More Accurate Optical Flow EstimationCode1
3D Object Detection with a Self-supervised Lidar Scene Flow BackboneCode1
DEFLOW: Self-supervised 3D Motion Estimation of Debris FlowCode1
Bi-PointFlowNet: Bidirectional Learning for Point Cloud Based Scene Flow EstimationCode1
MeteorNet: Deep Learning on Dynamic 3D Point Cloud SequencesCode1
AANet: Adaptive Aggregation Network for Efficient Stereo MatchingCode1
M-FUSE: Multi-frame Fusion for Scene Flow EstimationCode1
DifFlow3D: Toward Robust Uncertainty-Aware Scene Flow Estimation with Diffusion ModelCode1
Adversarial Self-Supervised Scene Flow EstimationCode1
DiffSF: Diffusion Models for Scene Flow EstimationCode1
CamLiFlow: Bidirectional Camera-LiDAR Fusion for Joint Optical Flow and Scene Flow EstimationCode1
RAFT-3D: Scene Flow using Rigid-Motion EmbeddingsCode1
GMSF: Global Matching Scene FlowCode1
Hidden Gems: 4D Radar Scene Flow Learning Using Cross-Modal SupervisionCode1
MambaFlow: A Novel and Flow-guided State Space Model for Scene Flow EstimationCode1
3DSFLabelling: Boosting 3D Scene Flow Estimation by Pseudo Auto-labellingCode1
Dual Task Learning by Leveraging Both Dense Correspondence and Mis-Correspondence for Robust Change Detection With Imperfect MatchesCode1
PointPWC-Net: A Coarse-to-Fine Network for Supervised and Self-Supervised Scene Flow Estimation on 3D Point CloudsCode1
PointPWC-Net: Cost Volume on Point Clouds for (Self-)Supervised Scene Flow EstimationCode1
A Large Dataset to Train Convolutional Networks for Disparity, Optical Flow, and Scene Flow EstimationCode1
Learning to Segment Rigid Motions from Two FramesCode1
Fast Kernel Scene FlowCode1
Fast Neural Scene FlowCode1
HPLFlowNet: Hierarchical Permutohedral Lattice FlowNet for Scene Flow Estimation on Large-scale Point CloudsCode1
FastPCI: Motion-Structure Guided Fast Point Cloud Frame InterpolationCode1
FESTA: Flow Estimation via Spatial-Temporal Attention for Scene Point CloudsCode1
FLOT: Scene Flow on Point Clouds Guided by Optimal TransportCode1
Flow4D: Leveraging 4D Voxel Network for LiDAR Scene Flow EstimationCode1
Deformation and Correspondence Aware Unsupervised Synthetic-to-Real Scene Flow Estimation for Point CloudsCode1
DELFlow: Dense Efficient Learning of Scene Flow for Large-Scale Point CloudsCode1
Neural Scene Flow PriorCode1
Show:102550
← PrevPage 1 of 4Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1FastNSFEPE 3-Way0.11Unverified
2FastFlow3DEPE 3-Way0.06Unverified
3NSFPEPE 3-Way0.06Unverified
4ZeroFlow 5x XLEPE 3-Way0.05Unverified
5SeFlowEPE 3-Way0.05Unverified
6TrackFlowEPE 3-Way0.05Unverified
7DeFlowEPE 3-Way0.03Unverified
#ModelMetricClaimedVerifiedStatus
1CamLiFlow (K)1px total85.31Unverified
2RAFT-3D (F)1px total78.82Unverified
3M-FUSE (K)1px total62.49Unverified
4CamLiFlow (F)1px total50.08Unverified
5RAFT-3D (K)1px total37.26Unverified
6M-FUSE (F)1px total34.9Unverified
#ModelMetricClaimedVerifiedStatus
1Self-Mono-SFSF-all49.54Unverified
2Multi-Mono-SFSF-all44.04Unverified
3PWOC-3DSF-all15.69Unverified
4CamLiRAFTSF-all4.26Unverified
#ModelMetricClaimedVerifiedStatus
1Self-Mono-SFD1-all31.25Unverified
2Multi-Mono-SFD1-all27.33Unverified
3EPCD1-all26.81Unverified
4EPC++D1-all23.84Unverified