AI Paper Atlas

Research map / Point clouds / Point cloud classification

Transferable 3D Adversarial Shape Completion using Diffusion Models

Xuelong Dai, Bin Xiao

ECCV 2024 · Point clouds › Point cloud classification · approach: Adversarial point cloud attacks · 3 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 ensemble adversarial guidance and model uncertainty to 3DAdvDiff3DAdvDiff · ShapeNet Chair, black-box transferaverage ASR (%): 19.0 -> 85.2Improved
add model uncertainty via random down-sampling (ablation)3DAdvDiff · ShapeNet Chairaverage black-box ASR (%): 19.0 -> 75.5Improved
add ensemble adversarial guidance (ablation)3DAdvDiff · ShapeNet Chairaverage black-box ASR (%): 19.0 -> 65.2Improved
increase number of random samples M (ablation)3DAdvDiff model uncertainty · ShapeNettransferability and running timeMixed
replace perturbation-based attack with diffusion shape completion (white-box 3DAdvDiff)PF-Attack / PGD baselines · ShapeNet All, black-box transferaverage black-box ASR (%): 6.2 -> 36.6Improved
integrate 3DAdvDiff boosting methods into PGD without diffusionPGD · ShapeNet whole test setaverage black-box ASR (%): 9.2 -> 25.0Improved
integrate 3DAdvDiff boosting methods into SI-AdvSI-Adv · ShapeNet whole test setPointNet ASR (white-box): 97.6 -> 70.5Hurt
use ensemble guidance with critical-point selection vs plain 3DAdvDiff3DAdvDiff · IF-Defense (PointNet)ASR (%): 30.0 -> 22.5Hurt
use ensemble version against defenses SRS, SOR, DUP-Net, Hybrid Training3DAdvDiff · Hybrid Training defenseASR (%): 9.4 -> 96.1Improved
add generation quality augmentation (l0 limit on critical points) (ablation)3DAdvDiff_ens · ShapeNetgeneration quality and ASRImproved
diffusion-based generation instead of perturbation attack3DAdvDiff_ens · ShapeNetrunning time (s): 1.1 -> 60.8Hurt

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