Research map / Point clouds / Shape and surface completion
PU-GAN: A Point Cloud Upsampling Adversarial Network
Ruihui Li, Xianzhi Li, Chi-Wing Fu, Daniel Cohen-Or, Pheng-Ann Heng
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 |
|---|---|---|---|
| add the discriminator (GAN framework) (ablation) | PU-GAN · Ablation test set | CD (1e-3): 0.57 -> 0.28 | Improved |
| add the discriminator (GAN framework) (ablation) | PU-GAN · Ablation test set | uniformity p=0.4% (1e-3): 7.02 -> 3.38 | Improved |
| add uniform loss (ablation) | PU-GAN · Ablation test set | CD (1e-3): 0.51 -> 0.28 | Improved |
| add uniform loss (ablation) | PU-GAN · Ablation test set | P2F (1e-3): 3.99 -> 2.33 | Improved |
| add self-attention unit (ablation) | PU-GAN · Ablation test set | CD (1e-3): 0.46 -> 0.28 | Improved |
| add self-attention unit (ablation) | PU-GAN · Ablation test set | HD (1e-3): 6.01 -> 4.64 | Improved |
| add up-down-up expansion unit (ablation) | PU-GAN · Ablation test set | CD (1e-3): 0.35 -> 0.28 | Improved |
| add up-down-up expansion unit (ablation) | PU-GAN · Ablation test set | P2F (1e-3): 3.02 -> 2.33 | Improved |
| add farthest sampling (ablation) | PU-GAN · Ablation test set | CD (1e-3): 0.31 -> 0.28 | Improved |
| add farthest sampling (ablation) | PU-GAN · Ablation test set | HD (1e-3): 4.96 -> 4.64 | Improved |
| use full PU-GAN instead of basic GAN baseline (ablation) | GAN baseline · Ablation test set | CD (1e-3): 0.59 -> 0.28 | Improved |
| upsample inputs with PU-GAN before classification | PointNet · ModelNet40 | classification accuracy (%): 82.4% -> 83.8% | Improved |
Later papers that compared with this method
7 later accepted papers report PU-GAN as a baseline in their results tables; 7 of them report a better number on most of the results they share with it.
- SPU-IMR: Self-supervised Arbitrary-scale Point Cloud Upsampling via Iterative Mask-recovery Network (AAAI 2025 · better on 2 of 3 shared results)
- RepKPU: Point Cloud Upsampling with Kernel Point Representation and Deformation (CVPR 2024 · better on 1 of 1 shared result)
- Deep Point Set Resampling via Gradient Fields (TPAMI 2022 · better on 3 of 3 shared results)
- PMP-Net++: Point Cloud Completion by Transformer-Enhanced Multi-Step Point Moving Paths (TPAMI 2022 · better on 2 of 2 shared results)
- PU-EVA: An Edge-Vector Based Approximation Solution for Flexible-Scale Point Cloud Upsampling (ICCV 2021 · better on 14 of 17 shared results)
- Point Cloud Upsampling via Disentangled Refinement (CVPR 2021 · better on 11 of 11 shared results)
- PU-GCN: Point Cloud Upsampling Using Graph Convolutional Networks (CVPR 2021 · better on 4 of 5 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
- PU-Net: Point Cloud Upsampling Network (CVPR 2018 · similarity 0.93)
- Grad-PU: Arbitrary-Scale Point Cloud Upsampling via Gradient Descent With Learned Distance Functions (CVPR 2023 · similarity 0.89)
- PU-EVA: An Edge-Vector Based Approximation Solution for Flexible-Scale Point Cloud Upsampling (ICCV 2021 · similarity 0.89)
- Arbitrary-Scale Point Cloud Upsampling by Voxel-Based Network with Latent Geometric-Consistent Learning (AAAI 2024 · similarity 0.88)
- Learning Continuous Implicit Field with Local Distance Indicator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2024 · similarity 0.87)
- SPU-IMR: Self-supervised Arbitrary-scale Point Cloud Upsampling via Iterative Mask-recovery Network (AAAI 2025 · similarity 0.87)
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.
