Research map / Point clouds / Registration and matching
SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registration
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use SpinNet's cylindrical, rotation-invariant descriptor instead of LMVD | 3DMatch feature matching · 3DMatch | FMR (%): 97.5 -> 97.6 | Helped |
| Use SpinNet's cylindrical, rotation-invariant descriptor instead of LMVD | Rotated 3DMatch feature matching · Rotated 3DMatch | FMR (%): 96.9 -> 97.5 | Helped |
| Train on 3DMatch and test with SpinNet rather than LMVD | Unseen ETH scenes · ETH, average across four scenes | FMR (%): 79.9 -> 92.8 | Helped |
| Use SpinNet rather than D3Feat-pred | KITTI odometry registration · KITTI odometry | Success (%): 99.81 -> 99.10 | Hurt |
| Train on KITTI and test SpinNet rather than D3Feat-pred | Rotated 3DMatch · 3DMatch (KITTI-trained; rotated) | FMR (%): 17.8 -> 84.2 | Helped |
| Train on 3DMatch and test SpinNet rather than D3Feat-pred | Unseen KITTI · KITTI (3DMatch-trained) | Success (%): 36.76 -> 81.44 | Helped |
| Add alignment with a reference axis (ablation) | SpinNet · 3DMatch (origin; τ₂ = 0.05) | FMR (%): 95.1 -> 97.6 | Helped |
| Add transformation of spherical voxels on the XY-plane (ablation) | SpinNet · 3DMatch (origin; τ₂ = 0.05) | FMR (%): 93.5 -> 97.6 | Helped |
| Use learned point-based voxel signatures instead of handcrafted point density (ablation) | SpinNet · ETH (origin; τ₂ = 0.05) | FMR (%): 42.6 -> 92.8 | Helped |
| Use 3DCCN instead of shared MLP layers to learn spatial context (ablation) | SpinNet · 3DMatch (origin; τ₂ = 0.05) | FMR (%): 64.4 -> 97.6 | Helped |
Closest papers by meaning
- Rethinking Rotation Invariance with Point Cloud Registration (AAAI 2023 · similarity 0.89)
- 3D Point Cloud Registration for Localization Using a Deep Neural Network Auto-Encoder (CVPR 2017 · similarity 0.89)
- Rotation-Invariant Local-to-Global Representation Learning for 3D Point Cloud (NeurIPS 2020 · similarity 0.89)
- Rotation-Invariant Transformer for Point Cloud Matching (CVPR 2023 · similarity 0.89)
- PosDiffNet: Positional Neural Diffusion for Point Cloud Registration in a Large Field of View with Perturbations (AAAI 2024 · similarity 0.87)
- TetraSphere: A Neural Descriptor for O(3)-Invariant Point Cloud Analysis (CVPR 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.