Research map / Point clouds / Point cloud classification
PointAugment: An Auto-Augmentation Framework for Point Cloud Classification
Ruihui Li, Xianzhi Li, Pheng-Ann Heng, Chi-Wing Fu
This paper tried a common change: Add dropout
Add dropout to displacement and transformation · ModelNet40 · accuracy: 92.5 -> 92.8 Improved
What 795 accepted papers report for this change
- 549 improved
- 87 no clear effect
- 12 mixed
- 147 hurt
- Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and Inference ICML 2026 · apply layer dropout to whole transformer blocks instead of attention and FFN sub-blocks separately · 20 TPP val loss, 271M · val loss: 2.377 -> 2.367Improved
- Quantifying the Uncertainty of Foundation Models with Singular Value Ensembles ICML 2026 · use MC Dropout · vision and SST-2 · accuracy: 86.3 -> 76.6 on Flowers102Hurt
- Motion Planning in Compressed Representation Spaces ICML 2026 · add nested dropout during training · WOMD guided generation · goal FDE: 4.145 -> 1.104Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 |
|---|---|---|---|
| Replace conventional data augmentation with PointAugment | Four classifiers · ModelNet40 | accuracy: 89.2 -> 90.9; 90.7 -> 92.9; 91.7 -> 92.7; 92.2 -> 93.4 | Improved |
| Replace conventional data augmentation with PointAugment | Four classifiers · ModelNet10 | accuracy: 91.9 -> 94.1; 93.3 -> 95.8; 94.2 -> 96.0; 94.8 -> 96.7 | Improved |
| Replace conventional data augmentation with PointAugment | Four classifiers · SHREC16 | accuracy: 84.4 -> 88.4; 85.1 -> 89.5; 86.6 -> 90.1; 87.0 -> 90.6 | Improved |
| Replace conventional data augmentation with PointAugment | Four classifiers · ModelNet40 | mAP: 70.5 -> 75.8; 81.3 -> 86.7; 83.2 -> 86.6; 85.3 -> 89.0 | Improved |
| Train with PointAugment rather than conventional DA | PointNet++ · ModelNet40 test corruption: jittering | accuracy: 90.3 -> 92.8 | Improved |
| Train with PointAugment rather than conventional DA | PointNet++ · ModelNet40 test corruption: scaling 0.9 | accuracy: 90.3 -> 92.8 | Improved |
| Train with PointAugment rather than conventional DA | PointNet++ · ModelNet40 test corruption: scaling 1.1 | accuracy: 90.3 -> 92.8 | Improved |
| Train with PointAugment rather than conventional DA | PointNet++ · ModelNet40 test corruption: rotation 90° | accuracy: 89.9 -> 92.7 | Improved |
| Train with PointAugment rather than conventional DA | PointNet++ · ModelNet40 test corruption: rotation 180° | accuracy: 89.7 -> 92.6 | Improved |
| Add point-wise displacement (ablation) | PointNet++ with conventional DA · ModelNet40 | accuracy: 90.7 -> 91.7 | Improved |
| Add shape-wise transformation (ablation) | PointNet++ with conventional DA · ModelNet40 | accuracy: 90.7 -> 91.9 | Improved |
| Add shape-wise transformation to point-wise displacement (ablation) | PointNet++ augmentor · ModelNet40 | accuracy: 91.7 -> 92.5 | Improved |
Closest papers by meaning
- Towards Robust Point Cloud Recognition With Sample-Adaptive Auto-Augmentation (TPAMI 2025 · similarity 0.88)
- Unsupervised Point Cloud Completion and Segmentation by Generative Adversarial Autoencoding Network (NeurIPS 2022 · similarity 0.87)
- Sample-adaptive Augmentation for Point Cloud Recognition Against Real-world Corruptions (ICCV 2023 · similarity 0.87)
- Point MixSwap: Attentional Point Cloud Mixing via Swapping Matched Structural Divisions (ECCV 2022 · similarity 0.87)
- Revisiting Point Cloud Shape Classification with a Simple and Effective Baseline (ICML 2021 · similarity 0.87)
- Rethinking Point Cloud Data Augmentation: Topologically Consistent Deformation (ICML 2025 · similarity 0.85)
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.
