AI Paper Atlas

Research map / Point clouds / Point cloud classification

Hiding Imperceptible Noise in Curvature-Aware Patches for 3D Point Cloud Attack

Mingyu Yang, Daizong Liu, Keke Tang, Pan Zhou, Lixing Chen, Junyang Chen

ECCV 2024 · Point clouds › Point cloud classification · approach: Adversarial point cloud attacks · 14 citations (OpenAlex)

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 changedWhereMeasuredEffect
add ISS critical-point subsampling instead of random point selection (ablation)WPA on PointNet · ModelNet40Hausdorff distance: 0.00695 -> 0.00437Improved
add ISS subsampling (ablation)WPA on PointNet · ShapeNetPartattack success rate: 87% -> 99% ASR; CD 0.00054 -> 0.00059Mixed
perturb highest-curvature patches along normal direction (HC) (ablation)WPA on PointNet · ModelNet40Hausdorff distance: 0.00437 -> 0.00197Improved
perturb lowest-curvature patches along tangent direction (LC) (ablation)WPA on PointNet · ModelNet40Hausdorff distance: 0.00437 -> 0.00204Improved
add curvature-specific patch perturbation on ShapeNetPart (ablation)WPA on PointNet · ShapeNetPartattack success rate: 99% -> 100%Improved
swap B-spline wavelet kernel for Meyer or Mexican Hat (ablation)WPA on PointNet · ModelNet40attack success rate and perturbation size: CD 0.0004-0.0005, ASR 100%No clear effect
perturb only a subset of points (proportion 0.5) with WPAWPA · ModelNet40perturbation sizeImproved
vary regularization weights beta1, beta2 (ablation)WPA · ModelNet40 PointNetattack success rate: 100% in all casesNo clear effect
apply WPA against SOR defense vs PGDPointNet · ModelNet40attack success rate: 52.3% -> 98.9%Improved
WPA against SRS defense versus SI-AdvPointNet · ModelNet40attack success rate: 85.6% vs 83.9%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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy