Research map / Point clouds / Point cloud classification
3DeformRS: Certifying Spatial Deformations on Point Clouds
Gabriel Pérez S., Juan C. Pérez, Motasem Alfarra, Silvio Giancola, Bernard Ghanem
Check 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 3DeformRS instead of 3DCertify for certification | PointNet · ModelNet40, 3° z-rotation, 1024 points | certified accuracy: 37.1 -> 100 | Improved |
| Use GPU acceleration for 3DeformRS | 3DeformRS · z-rotation certification | runtime (seconds per σ): 38.8 -> 7.3 | Improved |
| Add training augmentation with the target deformation | Point cloud DNNs · ModelNet40, deformations used for certification | certified accuracy | Improved |
| Add twisting training augmentation | Point cloud DNNs · ModelNet40, rotation and twisting certification | certified accuracy | Improved |
| Add training augmentations | PointNet++ and DGCNN · ModelNet40 | clean accuracy: PointNet++: +0.2%, -0.3%, -0.4%, -0.1%; DGCNN: +1.3%, +0.8%, +1.3%, -0.2% | Mixed |
| Use 3DeformRS rather than 3DCertify for certification | PointNet · ModelNet40, 3° z-rotation, 1024 points | certification runtime: ⇠89k seconds for 10°; ⇠40 seconds per σ | Improved |
| Increase point-cloud cardinality while using 3DeformRS | 3DeformRS with PointNet · ModelNet40, 3° z-rotation | certified accuracy: 98.8; 97.6; 97.8; 100; 100; 97.8; 100 | No clear effect |
Closest papers by meaning
- Robustness Certification for Point Cloud Models (ICCV 2021 · similarity 0.90)
- TPC: Transformation-Specific Smoothing for Point Cloud Models (ICML 2022 · similarity 0.88)
- Hide in Thicket: Generating Imperceptible and Rational Adversarial Perturbations on 3D Point Clouds (CVPR 2024 · similarity 0.87)
- ConcreTizer: Model Inversion Attack via Occupancy Classification and Dispersion Control for 3D Point Cloud Restoration (ICLR 2025 · similarity 0.87)
- Explicitly Perceiving and Preserving the Local Geometric Structures for 3D Point Cloud Attack (AAAI 2024 · similarity 0.87)
- DV-Matcher: Deformation-based Non-rigid Point Cloud Matching Guided by Pre-trained Visual Features (CVPR 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.
