mlsanityData to 14 Oct 2026

Lepard: Learning partial point cloud matching in rigid and deformable scenes

arXiv 2111.12591 ↗First posted 24 Nov 2021cs.CVcited by 85

Not on the map. Why?

Abstract (arXiv)

We present Lepard, a Learning based approach for partial point cloud matching in rigid and deformable scenes. The key characteristics are the following techniques that exploit 3D positional knowledge for point cloud matching: 1) An architecture that disentangles point cloud representation into feature space and 3D position space. 2) A position encoding method that explicitly reveals 3D relative distance information through the dot product of vectors. 3) A repositioning technique that modifies the crosspoint-cloud relative positions. Ablation studies demonstrate the effectiveness of the above techniques. In rigid cases, Lepard combined with RANSAC and ICP demonstrates state-of-the-art registration recall of 93.9% / 71.3% on the 3DMatch / 3DLoMatch. In deformable cases, Lepard achieves +27.1% / +34.8% higher non-rigid feature matching recall than the prior art on our newly constructed 4DMatch / 4DLoMatch benchmark.

Key facts. Cited by 85 papers in the map (data through 14 Oct 2026). Builds on 50 of its 86 references that are in the map. First posted on arXiv 24 Nov 2021.

Builds on

50 of 86 references.

References · 50

  1. 2017Attention Is All You Needcited by 60,361
  2. 2020End-to-End Object Detection with Transformerscited by 7,414
  3. 2014Fully Convolutional Networks for Semantic Segmentationcited by 4,859
  4. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  5. 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,443
  6. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,208
  7. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  8. 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 984
  9. 20163DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructionscited by 336
  10. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 321
  11. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 290
  12. 2021PoinTr: Diverse Point Cloud Completion with Geometry-Aware Transformerscited by 286
38 more
  1. 2018FlowNet3D: Learning Scene Flow in 3D Point Cloudscited by 242
  2. 2019PointNetLK: Robust & Efficient Point Cloud Registration using PointNetcited by 233
  3. 2020Deep Global Registrationcited by 229
  4. 2020RPM-Net: Robust Point Matching using Learned Featurescited by 182
  5. 2018PPFNet: Global Context Aware Local Features for Robust 3D Point Matchingcited by 181
  6. 2020D3Feat: Joint Learning of Dense Detection and Description of 3D Local Featurescited by 177
  7. 2018Neighbourhood Consensus Networkscited by 174
  8. 2018The Perfect Match: 3D Point Cloud Matching with Smoothed Densitiescited by 169
  9. 20183D-CODED : 3D Correspondences by Deep Deformationcited by 160
  10. 2016Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registrationcited by 159
  11. 2021CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registrationcited by 152
  12. 2019PRNet: Self-Supervised Learning for Partial-to-Partial Registrationcited by 146
  13. 2020SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registrationcited by 142
  14. 2021PointDSC: Robust Point Cloud Registration using Deep Spatial Consistencycited by 142
  15. 2018PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptorscited by 141
  16. 2020FLOT: Scene Flow on Point Clouds Guided by Optimal Transportcited by 129
  17. 2020Deep Geometric Functional Maps: Robust Feature Learning for Shape Correspondencecited by 125
  18. 20214DComplete: Non-Rigid Motion Estimation Beyond the Observable Surfacecited by 124
  19. 20193DRegNet: A Deep Neural Network for 3D Point Registrationcited by 96
  20. 2021Neural Scene Flow Priorcited by 91
  21. 2019DeepDeform: Learning Non-rigid RGB-D Reconstruction with Semi-supervised Datacited by 77
  22. 2016VolumeDeform: Real-time Volumetric Non-rigid Reconstructioncited by 75
  23. 2021You Only Hypothesize Once: Point Cloud Registration with Rotation-equivariant Descriptorscited by 68
  24. 2021Weakly Supervised Learning of Rigid 3D Scene Flowcited by 60
  25. 2021DPFM: Deep Partial Functional Mapscited by 58
  26. 2021OMNet: Learning Overlapping Mask for Partial-to-Partial Point Cloud Registrationcited by 51
  27. 2020Neural Non-Rigid Trackingcited by 49
  28. 2020End-to-End Learning Local Multi-view Descriptors for 3D Point Cloudscited by 47
  29. 2021MultiBodySync: Multi-Body Segmentation and Motion Estimation via 3D Scan Synchronizationcited by 46
  30. 2017Learning Compact Geometric Featurescited by 45
  31. 2021UnsupervisedR&R: Unsupervised Point Cloud Registration via Differentiable Renderingcited by 44
  32. 2019PointPWC-Net: A Coarse-to-Fine Network for Supervised and Self-Supervised Scene Flow Estimation on 3D Point Cloudscited by 38
  33. 2019SurfelWarp: Efficient Non-Volumetric Single View Dynamic Reconstructioncited by 31
  34. 2021Bootstrap Your Own Correspondencescited by 27
  35. 2021Self-supervised Geometric Perceptioncited by 26
  36. 2020Non-Rigid Puzzlescited by 18
  37. 2020Learning to Optimize Non-Rigid Trackingcited by 16
  38. 2020SplitFusion: Simultaneous Tracking and Mapping for Non-Rigid Scenescited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 6; 2023: 23; 2024: 30; 2025: 16; 2026: 10.

85 papers.

Sort:

By area:

  1. 20243D Diffuser Actor: Policy Diffusion with 3D Scene Representationscited by 282
  2. 2024Pixtral 12Bcited by 179
  3. 20233D Registration with Maximal Cliquescited by 84
  4. 2023GeoTransformer: Fast and Robust Point Cloud Registration with Geometric Transformercited by 84
  5. 2023Rotation-Invariant Transformer for Point Cloud Matchingcited by 69
  6. 2022Non-rigid Point Cloud Registration with Neural Deformation Pyramidcited by 42
  7. 2023Dynamic Point Fieldscited by 37
  8. 2023Robust Outlier Rejection for 3D Registration with Variational Bayescited by 31
  9. 2024MatchU: Matching Unseen Objects for 6D Pose Estimation from RGB-D Imagescited by 28
  10. 2022RIGA: Rotation-Invariant and Globally-Aware Descriptors for Point Cloud Registrationcited by 25
  11. 20233D Implicit Transporter for Temporally Consistent Keypoint Discoverycited by 20
  12. 2023Robust Multiview Point Cloud Registration with Reliable Pose Graph Initialization and History Reweightingcited by 19
38 more
  1. 2024One-Shot Imitation Learning with Invariance Matching for Robotic Manipulationcited by 18
  2. 2023Deep Graph-based Spatial Consistency for Robust Non-rigid Point Cloud Registrationcited by 15
  3. 2024Extend Your Own Correspondences: Unsupervised Distant Point Cloud Registration by Progressive Distance Extensioncited by 15
  4. 2022Fast and Robust Non-Rigid Registration Using Accelerated Majorization-Minimizationcited by 12
  5. 2022Learning Feature Descriptors for Pre- and Intra-operative Point Cloud Matching for Laparoscopic Liver Registrationcited by 12
  6. 2023Density-invariant Features for Distant Point Cloud Registrationcited by 12
  7. 2023Non-Rigid Shape Registration via Deep Functional Maps Priorcited by 12
  8. 2024SPAFormer: Sequential 3D Part Assembly with Transformerscited by 12
  9. 2024SRIF: Semantic Shape Registration Empowered by Diffusion-based Image Morphing and Flow Estimationcited by 12
  10. 2024Beyond Complete Shapes: A Benchmark for Quantitative Evaluation of 3D Shape Surface Matching Algorithmscited by 11
  11. 2025ACORD: An Expert-Annotated Retrieval Dataset for Legal Contract Draftingcited by 11
  12. 2024Correspondence-Free Non-Rigid Point Set Registration Using Unsupervised Clustering Analysiscited by 10
  13. 2024DV-Matcher: Deformation-based Non-Rigid Point Cloud Matching Guided by Pre-trained Visual Featurescited by 10
  14. 2023Learning Accurate Template Matching with Differentiable Coarse-to-Fine Correspondence Refinementcited by 9
  15. 2025Register Any Point: Scaling 3D Point Cloud Registration by Flow Matchingcited by 9
  16. 2023Root Pose Decomposition Towards Generic Non-rigid 3D Reconstruction with Monocular Videoscited by 8
  17. 2025Landmark-Free Preoperative-to-Intraoperative Registration in Laparoscopic Liver Resectioncited by 7
  18. 2022UTOPIC: Uncertainty-aware Overlap Prediction Network for Partial Point Cloud Registrationcited by 6
  19. 2024PointDifformer: Robust Point Cloud Registration With Neural Diffusion and Transformercited by 6
  20. 2025IC-World: In-Context Generation for Shared World Modelingcited by 6
  21. 2024Point Tree Transformer for Point Cloud Registrationcited by 5
  22. 2023Robust-DefReg: A Robust Deformable Point Cloud Registration Method based on Graph Convolutional Neural Networkscited by 4
  23. 2024ML-SemReg: Boosting Point Cloud Registration with Multi-level Semantic Consistencycited by 4
  24. 2024PointRegGPT: Boosting 3D Point Cloud Registration using Generative Point-Cloud Pairs for Trainingcited by 4
  25. 2024Resolving the Ambiguity of Complete-to-Partial Point Cloud Registration for Image-Guided Liver Surgery with Patches-to-Partial Matchingcited by 4
  26. 2025Occlusion-aware Non-Rigid Point Cloud Registration via Unsupervised Neural Deformation Correntropycited by 4
  27. 2024PosDiffNet: Positional Neural Diffusion for Point Cloud Registration in a Large Field of View with Perturbationscited by 3
  28. 2024Diff-Reg v1: Diffusion Matching Model for Registration Problemcited by 3
  29. 2024UMERegRobust - Universal Manifold Embedding Compatible Features for Robust Point Cloud Registrationcited by 3
  30. 2025FUSER: Feed-Forward MUltiview 3D Registration Transformer and SE(3)N Diffusion Refinementcited by 3
  31. 2026RegFormer++: An Efficient Large-Scale 3D LiDAR Point Registration Network with Projection-Aware 2D Transformercited by 3
  32. 2023Graph Matching Optimization Network for Point Cloud Registrationcited by 2
  33. 2023SphereNet: Learning a Noise-Robust and General Descriptor for Point Cloud Registrationcited by 2
  34. 2023SuPerPM: A Surgical Perception Framework Based on Deep Point Matching Learned from Physical Constrained Simulation Datacited by 2
  35. 2024Prim2Room: Layout-Controllable Room Mesh Generation from Primitivescited by 2
  36. 2024GERA: Geometric Embedding for Efficient Point Registration Analysiscited by 2
  37. 2024LoGDesc: Local geometric features aggregation for robust point cloud registrationcited by 2
  38. 20243D Registration in 30 Years: A Surveycited by 2

50 most cited of 85 All 85 (JSON)