mlsanityData to 7 Oct 2026

DetarNet: Decoupling Translation and Rotation by Siamese Network for Point Cloud Registration

arXiv 2112.14059 ↗First posted 28 Dec 2021cs.CVcited by 12

Not on the map. Why?

Abstract (arXiv)

Point cloud registration is a fundamental step for many tasks. In this paper, we propose a neural network named DetarNet to decouple the translation t and rotation R, so as to overcome the performance degradation due to their mutual interference in point cloud registration. First, a Siamese Network based Progressive and Coherent Feature Drift (PCFD) module is proposed to align the source and target points in high-dimensional feature space, and accurately recover translation from the alignment process. Then we propose a Consensus Encoding Unit (CEU) to construct more distinguishable features for a set of putative correspondences. After that, a Spatial and Channel Attention (SCA) block is adopted to build a classification network for finding good correspondences. Finally, the rotation is obtained by Singular Value Decomposition (SVD). In this way, the proposed network decouples the estimation of translation and rotation, resulting in better performance for both of them. Experimental results demonstrate that the proposed DetarNet improves registration performance on both indoor and outdoor scenes. Our code will be available in urlhttps://github.com/ZhiChen902/DetarNet.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Builds on 32 of its 57 references that are in the map. First posted on arXiv 28 Dec 2021.

Builds on

32 of 57 references.

References · 32

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  3. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  4. 2017Squeeze-and-Excitation Networkscited by 4,763
  5. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,201
  6. 2016Group Equivariant Convolutional Networkscited by 1,154
  7. 2018Open3D: A Modern Library for 3D Data Processingcited by 717
  8. 2015Multi-view Convolutional Neural Networks for 3D Shape Recognitioncited by 564
  9. 2018PVNet: Pixel-wise Voting Network for 6DoF Pose Estimationcited by 385
  10. 20163DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructionscited by 333
  11. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 318
  12. 2020PREDATOR: Registration of 3D Point Clouds with Low Overlapcited by 287
20 more
  1. 2018DeepIM: Deep Iterative Matching for 6D Pose Estimationcited by 283
  2. 2019PointNetLK: Robust & Efficient Point Cloud Registration using PointNetcited by 231
  3. 2020Deep Global Registrationcited by 228
  4. 2020RPM-Net: Robust Point Matching using Learned Featurescited by 181
  5. 2018PPFNet: Global Context Aware Local Features for Robust 3D Point Matchingcited by 180
  6. 2017Learning to Find Good Correspondencescited by 176
  7. 2020D3Feat: Joint Learning of Dense Detection and Description of 3D Local Featurescited by 175
  8. 2019Learning Two-View Correspondences and Geometry Using Order-Aware Networkcited by 148
  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. 2017Graph-Cut RANSACcited by 133
  12. 2019Neural-Guided RANSAC: Learning Where to Sample Model Hypothesescited by 129
  13. 20183DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registrationcited by 100
  14. 2018Neural Nearest Neighbors Networkscited by 98
  15. 20193DRegNet: A Deep Neural Network for 3D Point Registrationcited by 96
  16. 2021Deep Hough Voting for Robust Global Registrationcited by 53
  17. 2019NM-Net: Mining Reliable Neighbors for Robust Feature Correspondencescited by 29
  18. 2020Learnable Motion Coherence for Correspondence Pruningcited by 20
  19. 2018Learning and Matching Multi-View Descriptors for Registration of Point Cloudscited by 6
  20. 2021Cascade Network with Guided Loss and Hybrid Attention for Finding Good Correspondencescited by 2

Built on by

Cited by year

Citing papers per year: 2022: 4; 2023: 6; 2024: 1; 2025: 1.

12 papers.

  1. 2022SC^2-PCR: A Second Order Spatial Compatibility for Efficient and Robust Point Cloud Registrationcited by 62
  2. 2022Coarse-to-Fine Point Cloud Registration with SE(3)-Equivariant Representationscited by 9
  3. 2023PaRot: Patch-Wise Rotation-Invariant Network via Feature Disentanglement and Pose Restorationcited by 7
  4. 2022UTOPIC: Uncertainty-aware Overlap Prediction Network for Partial Point Cloud Registrationcited by 6
  5. 2023Efficient and Deterministic Search Strategy Based on Residual Projections for Point Cloud Registration with Correspondencescited by 5
  6. 2023Adaptive Multi-step Refinement Network for Robust Point Cloud Registrationcited by 4
  7. 2023Transformation Decoupling Strategy based on Screw Theory for Deterministic Point Cloud Registration with Gravity Priorcited by 3
  8. 2023DBDNet:Partial-to-Partial Point Cloud Registration with Dual Branches Decouplingcited by 2
  9. 20243D Registration in 30 Years: A Surveycited by 2
  10. 2023STDA-Meta: A Meta-Learning Framework for Few-Shot Traffic Predictioncited by 1
  11. 2022AS-PD: An Arbitrary-Size Downsampling Framework for Point Cloudscited by 0
  12. 2025Accelerating Outlier-robust Rotation Estimation by Stereographic Projectioncited by 0