Research map / Point clouds / Point cloud classification
SPE-Net: Boosting Point Cloud Analysis via Rotation Robustness Enhancement
Zhaofan Qiu, Yehao Li, Yu Wang, Yingwei Pan, Ting Yao, Tao Mei
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add encoding selection attention over CD, Z-RI, A-RI encodings, vs equal-weight combination (ablation) | SPE-MLP · ModelNet40 SPE-Net-S (Z/SO3) | accuracy: 26.0 -> 88.4 | Improved |
| add encoding selection attention vs equal-weight combination (ablation) | SPE-MLP · ModelNet40 SPE-Net-S (N/N) | accuracy: 93.5 -> 93.9 | Improved |
| add encoding selection attention vs equal-weight combination (ablation) | SPE-MLP · ModelNet40 SPE-Net-S (SO3/SO3) | accuracy: 90.7 -> 91.5 | Improved |
| use selection instead of A-RI only (ablation) | SPE-MLP · ModelNet40 SPE-Net-S (Z/SO3) | accuracy: 91.4 -> 88.4 | Hurt |
| use A-RI only vs CD only (ablation) | SPE-MLP · ModelNet40 SPE-Net-S (Z/SO3) | accuracy: 22.3 -> 91.4 | Improved |
| use A-RI only vs CD only (ablation) | SPE-MLP · ModelNet40 SPE-Net-S (N/N) | accuracy: 93.6 -> 90.6 | Hurt |
| increase mask-out epochs T (ablation) | SPE-Net · ModelNet40 | accuracy | Mixed |
| replace rotation-sensitive DGCNN with SPE-Net | DGCNN · ModelNet40 (Z/SO3) | accuracy: 20.6 -> 89.7 | Improved |
| replace PosPool with SPE-Net | PosPool · ModelNet40 (no rotation) | accuracy: 93.2 -> 94.0 | Improved |
| replace PosPool with SPE-Net | PosPool · S3DIS | mIoU: 66.7 -> 67.8 | Improved |
Closest papers by meaning
- PRIN/SPRIN: On Extracting Point-Wise Rotation Invariant Features (TPAMI 2021 · similarity 0.88)
- Deep Positional and Relational Feature Learning for Rotation-Invariant Point Cloud Analysis (ECCV 2020 · similarity 0.88)
- PARE-Net: Position-Aware Rotation-Equivariant Networks for Robust Point Cloud Registration (ECCV 2024 · similarity 0.88)
- CRIN: Rotation-Invariant Point Cloud Analysis and Rotation Estimation via Centrifugal Reference Frame (AAAI 2023 · similarity 0.87)
- PLATYPUS: Progressive Local Surface Estimator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2025 · similarity 0.85)
- RI-MAE: Rotation-Invariant Masked AutoEncoders for Self-Supervised Point Cloud Representation Learning (AAAI 2025 · similarity 0.85)
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.
