mlsanityData to 7 Oct 2026

Robust Multiview Point Cloud Registration with Reliable Pose Graph Initialization and History Reweighting

arXiv 2304.00467 ↗First posted 2 Apr 2023cs.CVcited by 17

Not on the map. Why?

Abstract (arXiv)

In this paper, we present a new method for the multiview registration of point cloud. Previous multiview registration methods rely on exhaustive pairwise registration to construct a densely-connected pose graph and apply Iteratively Reweighted Least Square (IRLS) on the pose graph to compute the scan poses. However, constructing a densely-connected graph is time-consuming and contains lots of outlier edges, which makes the subsequent IRLS struggle to find correct poses. To address the above problems, we first propose to use a neural network to estimate the overlap between scan pairs, which enables us to construct a sparse but reliable pose graph. Then, we design a novel history reweighting function in the IRLS scheme, which has strong robustness to outlier edges on the graph. In comparison with existing multiview registration methods, our method achieves 11% higher registration recall on the 3DMatch dataset and ~13% lower registration errors on the ScanNet dataset while reducing ~70% required pairwise registrations. Comprehensive ablation studies are conducted to demonstrate the effectiveness of our designs.

Key facts. Cited by 17 papers in the map (data through 7 Oct 2026). Builds on 39 of its 61 references that are in the map. First posted on arXiv 2 Apr 2023.

Builds on

39 of 61 references.

References · 39

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  3. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,088
  4. 2015NetVLAD: CNN architecture for weakly supervised place recognitioncited by 935
  5. 2019RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Cloudscited by 584
  6. 2019Deep Learning for 3D Point Clouds: A Surveycited by 557
  7. 2017Large-scale Point Cloud Semantic Segmentation with Superpoint Graphscited by 384
  8. 20163DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructionscited by 333
  9. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 318
  10. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 287
  11. 2019PointNetLK: Robust & Efficient Point Cloud Registration using PointNetcited by 231
  12. 2020Deep Global Registrationcited by 228
27 more
  1. 2022Geometric Transformer for Fast and Robust Point Cloud Registrationcited by 200
  2. 2020D3Feat: Joint Learning of Dense Detection and Description of 3D Local Featurescited by 175
  3. 2018The Perfect Match: 3D Point Cloud Matching with Smoothed Densitiescited by 169
  4. 2021CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registrationcited by 151
  5. 2019PRNet: Self-Supervised Learning for Partial-to-Partial Registrationcited by 146
  6. 2021PointDSC: Robust Point Cloud Registration using Deep Spatial Consistencycited by 142
  7. 2020SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registrationcited by 141
  8. 2020Feature-metric Registration: A Fast Semi-supervised Approach for Robust Point Cloud Registration without Correspondencescited by 108
  9. 20193DRegNet: A Deep Neural Network for 3D Point Registrationcited by 96
  10. 2021Robust Point Cloud Registration Framework Based on Deep Graph Matchingcited by 96
  11. 2022REGTR: End-to-end Point Cloud Correspondences with Transformerscited by 94
  12. 2020DeepGMR: Learning Latent Gaussian Mixture Models for Registrationcited by 91
  13. 2021Lepard: Learning partial point cloud matching in rigid and deformable scenescited by 85
  14. 2021A comprehensive survey on point cloud registrationcited by 84
  15. 2020Learning multiview 3D point cloud registrationcited by 78
  16. 2019DeepICP: An End-to-End Deep Neural Network for 3D Point Cloud Registrationcited by 77
  17. 2021You Only Hypothesize Once: Point Cloud Registration with Rotation-equivariant Descriptorscited by 68
  18. 2022SC^2-PCR: A Second Order Spatial Compatibility for Efficient and Robust Point Cloud Registrationcited by 62
  19. 2021Deep Hough Voting for Robust Global Registrationcited by 53
  20. 2019Learning Transformation Synchronizationcited by 29
  21. 2021HARA: A Hierarchical Approach for Robust Rotation Averagingcited by 22
  22. 2018Bayesian Pose Graph Optimization via Bingham Distributions and Tempered Geodesic MCMCcited by 19
  23. 2014A Solution for Multi-Alignment by Transformation Synchronisationcited by 15
  24. 2021Learning Iterative Robust Transformation Synchronizationcited by 13
  25. 2019Efficient and Robust Registration on the 3D Special Euclidean Groupcited by 6
  26. 2022Self-Supervised Correspondence Estimation via Multiview Registrationcited by 5
  27. 2020PointNetLK Revisitedcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 9; 2025: 5; 2026: 2.

17 papers.

Sort:

By area:

  1. 2023FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimatorscited by 29
  2. 2024Multiway Point Cloud Mosaicking with Diffusion and Global Optimizationcited by 14
  3. 2025Register Any Point: Scaling 3D Point Cloud Registration by Flow Matchingcited by 8
  4. 2024GeoNLF: Geometry guided Pose-Free Neural LiDAR Fieldscited by 6
  5. 2025FUSER: Feed-Forward MUltiview 3D Registration Transformer and SE(3)N Diffusion Refinementcited by 3
  6. 2024L-PR: Exploiting LiDAR Fiducial Marker for Unordered Low Overlap Multiview Point Cloud Registrationcited by 2
  7. 2024Incremental Multiview Point Cloud Registration with Two-stage Candidate Retrievalcited by 2
  8. 20243D Registration in 30 Years: A Surveycited by 2
  9. 2024Large-scale DSM registration via motion averagingcited by 1
  10. 2024Incremental Multiview Point Cloud Registrationcited by 1
  11. 2025Multiview Point Cloud Registration via Optimization in an Autoencoder Latent Spacecited by 1
  12. 2025Matching Distance and Geometric Distribution Aided Learning Multiview Point Cloud Registrationcited by 1
5 more
  1. 2025Correspondence-Free Multiview Point Cloud Registration via Depth-Guided Joint Optimisationcited by 1
  2. 2024Geometry Depth Consistency in RGBD Relative Pose Estimationcited by 0
  3. 2024Multiple Rotation Averaging with Constrained Reweighting Deep Matrix Factorizationcited by 0
  4. 2026Spectral-Geometric Neural Fields for Pose-Free LiDAR View Synthesiscited by 0
  5. 2026Spectral Consistency-Guided Multiview Point Cloud Registration for Low-Overlap Scenescited by 0