Research map / Point clouds / Point cloud classification
Certified L2-Norm Robustness of 3D Point Cloud Recognition in the Frequency Domain
Liang Zhou, Qiming Wang, Tianze Chen
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 FreqCert instead of randomized smoothing | certified defense (DGCNN) · ModelNet40 | certified radius: RS certifies nothing beyond eps=0.02; FreqCert up to eps=0.12 | Improved |
| use FreqCert instead of randomized smoothing | empirical defense · ModelNet40 (PGD) | empirical accuracy: RS collapses beyond eps=0.4; FreqCert over 70% at eps=1.0 | Improved |
| use FreqCert instead of IF-Defense under AdvPC | CurveNet · ModelNet40 (AdvPC eps=1.25) | empirical accuracy (%): 57.9 -> 73.3 | Improved |
| use FreqCert instead of DUP-Net under PGD | PointNet · ModelNet40 (PGD eps=1.25) | empirical accuracy (%): 74.8 -> 78.9 | Improved |
| increase number of slices m from 8 to 32 (ablation) | FreqCert · ModelNet40 | certified accuracy | Improved |
| increase number of slices m from 32 to 64 (ablation) | FreqCert · ModelNet40 | certified accuracy | No clear effect |
| use larger slices n=256 instead of n=32 (ablation) | FreqCert · ModelNet40 | certified accuracy: certificate vanishes near eps=0.08 (n=256) vs 0.14 (n=32) | Mixed |
| use DGCNN instead of PointNet backbone (ablation) | FreqCert · ModelNet40 | certified accuracy: gap typically below 2% | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DUP-Net — DUP-Net: Denoiser and Upsampler Network for 3D Adversarial Point Clouds Defense (ICCV 2019) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Improving Robustness of 3D Point Cloud Recognition from a Fourier Perspective (NeurIPS 2024 · similarity 0.92)
- PointCert: Point Cloud Classification With Deterministic Certified Robustness Guarantees (CVPR 2023 · similarity 0.91)
- Robustness Certification for Point Cloud Models (ICCV 2021 · similarity 0.90)
- PointGuard: Provably Robust 3D Point Cloud Classification (CVPR 2021 · similarity 0.89)
- Benchmarking and Analyzing Point Cloud Classification under Corruptions (ICML 2022 · similarity 0.88)
- Curvature-Invariant Adversarial Attacks for 3D Point Clouds (AAAI 2024 · 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.
