AI Paper Atlas

Research map / Point clouds / Point cloud classification

Sample-adaptive Augmentation for Point Cloud Recognition Against Real-world Corruptions

Jie Wang, Lihe Ding, Tingfa Xu, Shaocong Dong, Xinli Xu, Long Bai, Jianan Li

ICCV 2023 · Point clouds › Point cloud classification · approach: Adversarial point cloud attacks · 11 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 AdaptPoint augmentationRPC · ModelNet-CmCE (%): 86.3 -> 56.5Improved
add AdaptPoint augmentationRPC · ScanObjectNN-CmCE (%): 132.6 -> 96.5Improved
add AdaptPoint augmentationPoint-MAE · ShapeNet-C part segmentationmCE (%): 92.7 -> 76.9Improved
add AdaptPoint augmentationDGCNN · ScanObjectNN cleanOA (%): 85.8 -> 84.4Hurt
add feedback loss (ablation)AdaptPoint (PointNeXt) · ScanObjectNN-CmCE (%): 82.1 -> 78.3Improved
add adversarial discriminator loss (ablation)AdaptPoint (PointNeXt) · ScanObjectNN-CmCE (%): 83.2 -> 78.3Improved
add both deformation and mask controllers (ablation)AdaptPoint (PointNeXt) · ScanObjectNN-CmCE (%): 92.1 -> 78.3Improved
add mask controller to deformation only (ablation)AdaptPoint (PointNeXt) · ScanObjectNN-CmCE (%): 86.4 -> 78.3Improved
increase anchors from 4 to 16 (ablation)AdaptPoint deformation controller · ScanObjectNN-CmCE (%): 78.3 -> 81.4Hurt
change feedback loss weight lambda from 1 to 0.5 or 2 (ablation)AdaptPoint (PointNeXt) · ScanObjectNN-CmCE (%): 78.3 -> 78.7 (0.5), 79.0 (2)Hurt
apply AdaptPoint on PointNet++ for attack defensePointNet++ · Point cloud attack defense Drop-200OA (%): 79.09 (IF-Defense) -> 83.59Improved
apply AdaptPoint instead of IF-DefensePointNet++ · Point cloud attack defense kNNOA (%): 85.62 -> 77.67Hurt

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