Research map / Point clouds / Point cloud classification
Rethinking the compositionality of point clouds through regularization in the hyperbolic space
Antonio Montanaro, Diego Valsesia, Enrico Magli
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 HyCoRe regularizer | PointMLP · ModelNet40 | OA (%): 93.4 -> 94.3 | Improved |
| add HyCoRe regularizer | PointMLP · ScanObjectNN | OA (%): 86.1 -> 87.2 | Improved |
| add HyCoRe regularizer | PointNeXt · ScanObjectNN | OA (%): 88.0 -> 88.3 | Improved |
| add hyperbolic hierarchy regularizer R_hier (ablation) | DGCNN · ScanObjectNN | AA (%): 77.8 -> 77.9 | No clear effect |
| add hyperbolic contrastive regularizer R_contr (ablation) | DGCNN · ScanObjectNN | AA (%): 77.8 -> 79.2 | Improved |
| combine both regularizers instead of R_contr alone (ablation) | DGCNN · ScanObjectNN | AA (%): 79.2 -> 80.2 | Improved |
| use hyperbolic space instead of Euclidean for regularizer (HyCoRe vs EuCoRe) (ablation) | DGCNN · ScanObjectNN | AA (%): 79.0 -> 80.2 | Improved |
| add regularizer to hyperbolic head (Hype-DGCNN -> HyCoRe) (ablation) | Hype-DGCNN · ScanObjectNN | AA (%): 76.5 -> 80.2 | Improved |
| lower curvature from 1 to 0.01 (ablation) | DGCNN+HyCoRe · ScanObjectNN | AA (%): 80.2 -> 78.5 | Hurt |
| replace Euclidean head with unregularized hyperbolic head (ablation) | DGCNN · ScanObjectNN | AA (%): 77.8 -> 76.5 | Hurt |
Closest papers by meaning
- PointGLR: Unsupervised Structural Representation Learning of 3D Point Clouds (TPAMI 2022 · similarity 0.88)
- Hyperbolic Uncertainty-Aware Few-Shot Incremental Point Cloud Segmentation (CVPR 2025 · similarity 0.88)
- BoW Pooling: A Plug-and-Play Unit for Feature Aggregation of Point Clouds (AAAI 2021 · similarity 0.88)
- Global-Local Bidirectional Reasoning for Unsupervised Representation Learning of 3D Point Clouds (CVPR 2020 · similarity 0.88)
- An Explicit Frame Construction for Normalizing 3D Point Clouds (ICML 2024 · similarity 0.87)
- Dual Manifold Regularization Steered Robust Representation Learning for Point Cloud Analysis (AAAI 2025 · similarity 0.87)
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.
