Research map / Point clouds / Point cloud classification
TearingNet: Point Cloud Autoencoder To Learn Topology-Friendly Representations
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 |
|---|---|---|---|
| Add iterative topology tearing, using a trial folding before the Tearing network (ablation) | TearingNet · KIMO-3 through KIMO-6 | EMD: 1.055 -> 0.940; 1.787 -> 1.438; 2.565 -> 1.872; 3.476 -> 2.614 | Helped |
| Add a trial Folding network before the Tearing network (ablation) | TearingNet · KIMO-3 through KIMO-6 | CD (×10^-2): 5.05 -> 4.78; 6.58 -> 6.43; 8.55 -> 8.29; 10.86 -> 10.23 | Mixed |
| Add graph filtering after the second Folding network (ablation) | TearingNet · KIMO-3 through KIMO-6 | EMD: 0.958 -> 0.940; 1.441 -> 1.438; 1.879 -> 1.872; 2.648 -> 2.614 | Helped |
| Add a third iteration of the Tearing and Folding networks (ablation) | TearingNet · KIMO-3 through KIMO-6 | EMD: 0.940 -> 0.941; 1.438 -> 1.361; 1.872 -> 1.867; 2.614 -> 2.315 | Mixed |
| Use topology tearing rather than a genus-zero primitive decoder | TearingNet · KIMO-3 through KIMO-6 | object-counting MAE (×10^-1): 0.204 -> 0.123; 3.031 -> 1.721; 6.340 -> 5.079; 8.527 -> 8.026 | Helped |
| Use topology tearing rather than a genus-zero primitive decoder | TearingNet · KIMO-4 through KIMO-6 | pedestrian detection accuracy (%): 80.12 -> 83.42; 77.10 -> 79.74; 82.92 -> 84.55 | Helped |
Closest papers by meaning
- FoldingNet: Point Cloud Auto-Encoder via Deep Grid Deformation (CVPR 2018 · similarity 0.89)
- Density-Preserving Deep Point Cloud Compression (CVPR 2022 · similarity 0.89)
- LinNet: Linear Network for Efficient Point Cloud Representation Learning (NeurIPS 2024 · similarity 0.87)
- 3D Feature Prediction for Masked-AutoEncoder-Based Point Cloud Pretraining (ICLR 2024 · similarity 0.87)
- Rethinking Point Cloud Data Augmentation: Topologically Consistent Deformation (ICML 2025 · similarity 0.86)
- SIEVE: Overcoming Topological Obstruction in Equivariant Self-Supervised Learning (NeurIPS 2026 · 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.