Research map / Point clouds / Point cloud classification
E2PN: Efficient SE(3)-Equivariant Point Network
Read the paperCheck 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 E2PN instead of EPN | SE(3)-equivariant point-cloud classification · ModelNet40, SO(3) train/test, Noisy | accuracy (%): 84.63 -> 86.99 | Helped |
| Use E2PN instead of EPN | SE(3)-equivariant point-cloud classification · ModelNet40, Id train/SO(3) test, Noisy | accuracy (%): 30.99 -> 39.54 | Helped |
| Use E2PN instead of EPN | SE(3)-equivariant point-cloud classification · ModelNet40, Id train/Id test, Noisy | accuracy (%): 90.03 -> 90.66 | Mixed |
| Use E2PN instead of EPN | rotation estimation · ModelNet40 pose estimation | mean angular error (°): 1.10 -> 1.20 | No clear effect |
| Use E2PN instead of EPN | point-cloud keypoint descriptors · 3DMatch keypoint matching, Average | average recall (%): 97.6 -> 97.3 | Hurt |
| Add the permutation layer instead of using GA pooling (ablation) | E2PN · ModelNet40 classification, SO(3) train/test, Noisy | accuracy (%): 85.04 -> 86.99 | Helped |
| Add symmetric kernels and their efficient feature gathering (ablation) | E2PN · ModelNet40 classification | speed (fps), train/test: 7.8 / 7.8 -> 9.1 / 10.3 | Helped |
| Add symmetric kernels and their efficient feature gathering (ablation) | E2PN · 3DMatch keypoint matching | speed (fps), train/test: 2.6 / 16.7 -> 3.7 / 23.6 | Helped |
Closest papers by meaning
- Learning Coordinate-based Convolutional Kernels for Continuous SE(3) Equivariant and Efficient Point Cloud Analysis (CVPR 2026 · similarity 0.94)
- Equivariant Point Network for 3D Point Cloud Analysis (CVPR 2021 · similarity 0.94)
- Fast, Expressive $\mathrm{SE}(n)$ Equivariant Networks through Weight-Sharing in Position-Orientation Space (ICLR 2024 · similarity 0.90)
- SIEVE: Overcoming Topological Obstruction in Equivariant Self-Supervised Learning (NeurIPS 2026 · similarity 0.88)
- Pose-Transformed Equivariant Network for 3D Point Trajectory Prediction (CVPR 2024 · similarity 0.88)
- Approximately Piecewise E(3) Equivariant Point Networks (ICLR 2024 · similarity 0.88)
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.