mlsanityData to 14 Oct 2026

BUFFER-X: Towards Zero-Shot Point Cloud Registration in Diverse Scenes

arXiv 2503.07940 ↗First posted 11 Mar 2025cs.CV cs.RO eess.IVcited by 8

Not on the map. Why?

Abstract (arXiv)

Recent advances in deep learning-based point cloud registration have improved generalization, yet most methods still require retraining or manual parameter tuning for each new environment. In this paper, we identify three key factors limiting generalization: (a) reliance on environment-specific voxel size and search radius, (b) poor out-of-domain robustness of learning-based keypoint detectors, and (c) raw coordinate usage, which exacerbates scale discrepancies. To address these issues, we present a zero-shot registration pipeline called BUFFER-X by (a) adaptively determining voxel size/search radii, (b) using farthest point sampling to bypass learned detectors, and (c) leveraging patch-wise scale normalization for consistent coordinate bounds. In particular, we present a multi-scale patch-based descriptor generation and a hierarchical inlier search across scales to improve robustness in diverse scenes. We also propose a novel generalizability benchmark using 11 datasets that cover various indoor/outdoor scenarios and sensor modalities, demonstrating that BUFFER-X achieves substantial generalization without prior information or manual parameter tuning for the test datasets. Our code is available at https://github.com/MIT-SPARK/BUFFER-X.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 53 of its 97 references that are in the map. First posted on arXiv 11 Mar 2025.

Builds on

53 of 97 references.

References · 53

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  3. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,451
  4. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,208
  5. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  6. 2023ScanNet++: A High-Fidelity Dataset of 3D Indoor Scenescited by 764
  7. 2023Point Transformer V3: Simpler, Faster, Strongercited by 589
  8. 2020Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR Segmentationcited by 389
  9. 2022Point Transformer V2: Grouped Vector Attention and Partition-based Poolingcited by 388
  10. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 290
  11. 2020TEASER: Fast and Certifiable Point Cloud Registrationcited by 274
  12. 2017A General and Adaptive Robust Loss Functioncited by 236
38 more
  1. 2020Deep Global Registrationcited by 229
  2. 2020D3Feat: Joint Learning of Dense Detection and Description of 3D Local Featurescited by 177
  3. 2016Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registrationcited by 159
  4. 2021CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registrationcited by 152
  5. 2020SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registrationcited by 142
  6. 2022KISS-ICP: In Defense of Point-to-Point ICP -- Simple, Accurate, and Robust Registration If Done the Right Waycited by 132
  7. 20183DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registrationcited by 103
  8. 2019Graduated Non-Convexity for Robust Spatial Perception: From Non-Minimal Solvers to Global Outlier Rejectioncited by 94
  9. 20233D Registration with Maximal Cliquescited by 84
  10. 2023GeoTransformer: Fast and Robust Point Cloud Registration with Geometric Transformercited by 84
  11. 2021Only Train Once: A One-Shot Neural Network Training And Pruning Frameworkcited by 70
  12. 2021LCDNet: Deep Loop Closure Detection and Point Cloud Registration for LiDAR SLAMcited by 67
  13. 2020The Newer College Dataset: Handheld LiDAR, Inertial and Vision with Ground Truthcited by 61
  14. 2021Learning general and distinctive 3D local deep descriptors for point cloud registrationcited by 42
  15. 2023A Survey on Global LiDAR Localization: Challenges, Advances and Open Problemscited by 40
  16. 2019A Polynomial-time Solution for Robust Registration with Extreme Outlier Ratescited by 38
  17. 2021Patchwork: Concentric Zone-based Region-wise Ground Segmentation with Ground Likelihood Estimation Using a 3D LiDAR Sensorcited by 29
  18. 2020Distinctive 3D local deep descriptorscited by 21
  19. 2022A Single Correspondence Is Enough: Robust Global Registration to Avoid Degeneracy in Urban Environmentscited by 19
  20. 2020Kimera-Multi: a System for Distributed Multi-Robot Metric-Semantic Simultaneous Localization and Mappingcited by 18
  21. 2023Quatro++: Robust Global Registration Exploiting Ground Segmentation for Loop Closing in LiDAR SLAMcited by 18
  22. 2024PARE-Net: Position-Aware Rotation-Equivariant Networks for Robust Point Cloud Registrationcited by 18
  23. 2023Instant Domain Augmentation for LiDAR Semantic Segmentationcited by 16
  24. 2023Resilient and Distributed Multi-Robot Visual SLAM: Datasets, Experiments, and Lessons Learnedcited by 16
  25. 2024FastMAC: Stochastic Spectral Sampling of Correspondence Graphcited by 16
  26. 2024Scalable 3D Registration via Truncated Entry-wise Absolute Residualscited by 16
  27. 2021PHASER: a Robust and Correspondence-free Global Pointcloud Registrationcited by 15
  28. 2024Extend Your Own Correspondences: Unsupervised Distant Point Cloud Registration by Progressive Distance Extensioncited by 15
  29. 2023Accelerating Globally Optimal Consensus Maximization in Geometric Visioncited by 14
  30. 2021TriVoC: Efficient Voting-based Consensus Maximization for Robust Point Cloud Registration with Extreme Outlier Ratioscited by 11
  31. 2023Segregator: Global Point Cloud Registration with Semantic and Geometric Cuescited by 11
  32. 2024Correspondence-Free Non-Rigid Point Set Registration Using Unsupervised Clustering Analysiscited by 10
  33. 2021PointPCA: Point Cloud Objective Quality Assessment Using PCA-Based Descriptorscited by 9
  34. 2022Multi-Modal Lidar Dataset for Benchmarking General-Purpose Localization and Mapping Algorithmscited by 9
  35. 20233D-BBS: Global Localization for 3D Point Cloud Scan Matching Using Branch-and-Bound Algorithmcited by 8
  36. 2024Learning Instance-Aware Correspondences for Robust Multi-Instance Point Cloud Registration in Cluttered Scenescited by 7
  37. 2024RANSAC Back to SOTA: A Two-stage Consensus Filtering for Real-time 3D Registrationcited by 7
  38. 2022Learning to Register Unbalanced Point Pairscited by 5

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 6.

8 papers.

  1. 2025Register Any Point: Scaling 3D Point Cloud Registration by Flow Matchingcited by 9
  2. 2025RARE: Refine Any Registration of Pairwise Point Clouds via Zero-Shot Learningcited by 5
  3. 2026Toward Zero-Shot Point Cloud Registration Across Diverse Scales, Scenes, and Sensor Setupscited by 1
  4. 2026Long-Term Multi-Session 3D Reconstruction Under Substantial Appearance Changecited by 0
  5. 2026Paired-CSLiDAR: Height-Stratified Registration for Cross-Source Aerial-Ground LiDAR Pose Refinementcited by 0
  6. 2026PROSE: Training-Free Egocentric Scene Registration with Vision-Language Modelscited by 0
  7. 2026CVSD-Reg: Cross-Modal Visual Semantic Prior Distillation for Robust LiDAR Registrationcited by 0
  8. 2026BridgeMatch: Conditional Transport Bridges in Matching Matrix Space for 3D Deformable Registrationcited by 0