Research map / Point clouds / Point cloud classification
Attention-Based Point Cloud Edge Sampling
Chengzhi Wu, Junwei Zheng, Julius Pfrommer, Jürgen Beyerer
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 |
|---|---|---|---|
| Use global-based APES instead of local-based APES | APES classification model · ModelNet40 | overall accuracy: 93.5% -> 93.8% | Improved |
| Use APES global sampling rather than the strongest reported baseline sampling method | PointNet sampling evaluation · ModelNet40 classification, 32 points | accuracy: 86.18 -> 88.56 | Improved |
| Use N2P attention instead of EdgeConv for feature learning (ablation) | APES local · ModelNet40 | overall accuracy: 93.02 -> 93.47 | Improved |
| Use N2P attention instead of P2P attention for feature learning (ablation) | APES local · ModelNet40 | overall accuracy: 93.30 -> 93.47 | Improved |
| Use N2P attention instead of EdgeConv for feature learning (ablation) | APES global · ModelNet40 | overall accuracy: 93.18 -> 93.81 | Improved |
| Use N2P attention instead of P2P attention for feature learning (ablation) | APES global · ModelNet40 | overall accuracy: 93.46 -> 93.81 | Improved |
| Increase embedding dimension from 64 to 128 (ablation) | APES local · ModelNet40 | overall accuracy: 93.10 -> 93.47 | Improved |
| Increase embedding dimension from 128 to 192 (ablation) | APES local · ModelNet40 | overall accuracy: 93.47 -> 93.54 | Improved |
| Increase embedding dimension from 64 to 128 (ablation) | APES global · ModelNet40 | overall accuracy: 93.34 -> 93.81 | Improved |
| Increase embedding dimension from 128 to 192 (ablation) | APES global · ModelNet40 | overall accuracy: 93.81 -> 93.83 | Improved |
| Increase local neighborhood size from k=8 to k=32 (ablation) | APES local · ModelNet40 | overall accuracy: 93.14 -> 93.47 | Improved |
| Increase local neighborhood size from k=32 to k=512 (ablation) | APES local · ModelNet40 | overall accuracy: 93.47 -> 93.63 | Improved |
Closest papers by meaning
- SAMBLE: Shape-Specific Point Cloud Sampling for an Optimal Trade-Off Between Local Detail and Global Uniformity (CVPR 2025 · similarity 0.88)
- LTA-PCS: Learnable Task-Agnostic Point Cloud Sampling (CVPR 2024 · similarity 0.87)
- GCD-Sampling: A General Cross-scale Decoupled Sampling for Point Cloud (AAAI 2025 · similarity 0.86)
- MICAS: Multi-grained In-Context Adaptive Sampling for 3D Point Cloud Processing (CVPR 2025 · similarity 0.85)
- Point Cloud Pre-training with Diffusion Models (CVPR 2024 · similarity 0.84)
- Point Cloud Upsampling Using Conditional Diffusion Module with Adaptive Noise Suppression (CVPR 2025 · similarity 0.84)
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.
