mlsanityData to 7 Oct 2026

Global Point Cloud Registration Network for Large Transformations

arXiv 2403.18040 ↗First posted 26 Mar 2024cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Three-dimensional data registration is an established yet challenging problem that is key in many different applications, such as mapping the environment for autonomous vehicles, and modeling objects and people for avatar creation, among many others. Registration refers to the process of mapping multiple data into the same coordinate system by means of matching correspondences and transformation estimation. Novel proposals exploit the benefits of deep learning architectures for this purpose, as they learn the best features for the data, providing better matches and hence results. However, the state of the art is usually focused on cases of relatively small transformations, although in certain applications and in a real and practical environment, large transformations are very common. In this paper, we present ReLaTo (Registration for Large Transformations), an architecture that faces the cases where large transformations happen while maintaining good performance for local transformations. This proposal uses a novel Softmax pooling layer to find correspondences in a bilateral consensus manner between two point sets, sampling the most confident matches. These matches are used to estimate a coarse and global registration using weighted Singular Value Decomposition (SVD). A target-guided denoising step is then applied to both the obtained matches and latent features, estimating the final fine registration considering the local geometry. All these steps are carried out following an end-to-end approach, which has been shown to improve 10 state-of-the-art registration methods in two datasets commonly used for this task (ModelNet40 and KITTI), especially in the case of large transformations.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 29 of its 64 references that are in the map. First posted on arXiv 26 Mar 2024.

Builds on

29 of 64 references.

References · 29

  1. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  2. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  3. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 318
  4. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 287
  5. 2020Deep Global Registrationcited by 228
  6. 2020RPM-Net: Robust Point Matching using Learned Featurescited by 181
  7. 2018PPFNet: Global Context Aware Local Features for Robust 3D Point Matchingcited by 180
  8. 2021CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registrationcited by 151
  9. 2019PRNet: Self-Supervised Learning for Partial-to-Partial Registrationcited by 146
  10. 2021PointDSC: Robust Point Cloud Registration using Deep Spatial Consistencycited by 142
  11. 2020SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registrationcited by 141
  12. 2018PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptorscited by 140
17 more
  1. 2022Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Cloudscited by 131
  2. 20193DRegNet: A Deep Neural Network for 3D Point Registrationcited by 96
  3. 2021Robust Point Cloud Registration Framework Based on Deep Graph Matchingcited by 96
  4. 2022REGTR: End-to-end Point Cloud Correspondences with Transformerscited by 94
  5. 2020DeepGMR: Learning Latent Gaussian Mixture Models for Registrationcited by 91
  6. 2023GeoTransformer: Fast and Robust Point Cloud Registration with Geometric Transformercited by 83
  7. 2020FPConv: Learning Local Flattening for Point Convolutioncited by 55
  8. 2021OMNet: Learning Overlapping Mask for Partial-to-Partial Point Cloud Registrationcited by 50
  9. 2021HRegNet: A Hierarchical Network for Large-scale Outdoor LiDAR Point Cloud Registrationcited by 47
  10. 2022Generalized Differentiable RANSACcited by 30
  11. 2021PCAM: Product of Cross-Attention Matrices for Rigid Registration of Point Cloudscited by 20
  12. 2021Deep Weighted Consensus: Dense correspondence confidence maps for 3D shape registrationcited by 5
  13. 2020When Deep Learning Meets Data Alignment: A Review on Deep Registration Networks (DRNs)cited by 4
  14. 20183D non-rigid registration using color: Color Coherent Point Driftcited by 3
  15. 2018DUGMA: Dynamic Uncertainty-Based Gaussian Mixture Alignmentcited by 3
  16. 2017μ-MAR: Multiplane 3D Marker based Registration for Depth-sensing Camerascited by 1
  17. 2021Two Heads are Better than One: Geometric-Latent Attention for Point Cloud Classification and Segmentationcited by 1

Built on by

Not cited yet.