mlsanityData to 7 Oct 2026

Deep Weighted Consensus: Dense correspondence confidence maps for 3D shape registration

arXiv 2105.02714 ↗First posted 6 May 2021cs.CV cs.AIcited by 5

Not on the map. Why?

Abstract (arXiv)

We present a new paradigm for rigid alignment between point clouds based on learnable weighted consensus which is robust to noise as well as the full spectrum of the rotation group. Current models, learnable or axiomatic, work well for constrained orientations and limited noise levels, usually by an end-to-end learner or an iterative scheme. However, real-world tasks require us to deal with large rotations as well as outliers and all known models fail to deliver. Here we present a different direction. We claim that we can align point clouds out of sampled matched points according to confidence level derived from a dense, soft alignment map. The pipeline is differentiable, and converges under large rotations in the full spectrum of SO(3), even with high noise levels. We compared the network to recently presented methods such as DCP, PointNetLK, RPM-Net, PRnet, and axiomatic methods such as ICP and Go-ICP. We report here a fundamental boost in performance.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 18 of its 42 references that are in the map. First posted on arXiv 6 May 2021.

Builds on

18 of 42 references.

References · 18

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 2016Layer Normalizationcited by 5,579
  3. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  4. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
  5. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,220
  6. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,202
  7. 2015Convolutional Networks on Graphs for Learning Molecular Fingerprintscited by 751
  8. 2015Empirical Evaluation of Rectified Activations in Convolutional Networkcited by 515
  9. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 318
  10. 2019PointNetLK: Robust & Efficient Point Cloud Registration using PointNetcited by 231
  11. 2020Deep Global Registrationcited by 228
  12. 2018SO-Net: Self-Organizing Network for Point Cloud Analysiscited by 210
6 more
  1. 2020RPM-Net: Robust Point Matching using Learned Featurescited by 181
  2. 2018PPFNet: Global Context Aware Local Features for Robust 3D Point Matchingcited by 180
  3. 2016Go-ICP: A Globally Optimal Solution to 3D ICP Point-Set Registrationcited by 159
  4. 2019PRNet: Self-Supervised Learning for Partial-to-Partial Registrationcited by 146
  5. 2019Rotation Invariant Convolutions for 3D Point Clouds Deep Learningcited by 80
  6. 2020A Rotation-Invariant Framework for Deep Point Cloud Analysiscited by 62

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 1; 2023: 0; 2024: 2; 2025: 0; 2026: 1.

5 papers.

  1. 2022Deep Confidence Guided Distance for 3D Partial Shape Registrationcited by 2
  2. 2024Equivariant Local Reference Frames for Unsupervised Non-rigid Point Cloud Shape Correspondencecited by 2
  3. 2021DPC: Unsupervised Deep Point Correspondence via Cross and Self Constructioncited by 0
  4. 2024Global Point Cloud Registration Network for Large Transformationscited by 0
  5. 2026R-SLPR: Region-based Small-to-Large Point-cloud Registration with Contrastive Learningcited by 0