Research map / Point clouds / Shape and surface completion
PUGeo-Net: A Geometry-centric Network for 3D Point Cloud Upsampling
Yue Qian, Junhui Hou, Sam Kwong, Ying He
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 feature recalibration (ablation) | PUGeo-Net · Ablation (R=8 as stated) | CD (x10^-2): 0.325 -> 0.323 | Improved |
| add normal prediction and supervision (ablation) | PUGeo-Net · Ablation (R=8 as stated) | CD (x10^-2): 0.331 -> 0.323; HD 2.232 -> 1.011 | Improved |
| add learned adaptive 2D sampling (ablation) | PUGeo-Net · Ablation (R=8 as stated) | CD (x10^-2): 0.326 -> 0.323 | Improved |
| predict linear transformation instead of direct MLP regression (ablation) | PUGeo-Net · Ablation (R=8 as stated) | CD (x10^-2): 0.394 -> 0.323; CD# 1.855 -> 0.524 | Improved |
| add coarse-to-fine prediction (ablation) | PUGeo-Net · Ablation (R=8 as stated) | CD (x10^-2): CD 0.330 -> 0.323; HD# 5.241 -> 5.267 | Mixed |
| use PUGeo-Net predicted normals instead of PCA with 35 neighbours | PUGeo-Net · Normal estimation (R=8) | CD# (x10^-2): 0.553 -> 0.549 | Improved |
| use PUGeo-Net instead of PU-Net-M (PU-Net predicting 6 outputs) | PUGeo-Net · Normal estimation (R=8) | CD# (x10^-2): 0.678 -> 0.549 | Improved |
| vary PCA neighbourhood size 10 to 35 | PCA normals · Normal estimation (R=8) | CD# (x10^-2): 0.586 -> 0.553 | Mixed |
| use PUGeo-Net instead of MPU at 4x | PUGeo-Net · Upsampling 4x | CD (x10^-2): 0.573 -> 0.558 | Improved |
| use PUGeo-Net instead of MPU at 16x | PUGeo-Net · Upsampling 16x | CD (x10^-2): 0.344 -> 0.323 | Improved |
Later papers that compared with this method
1 later accepted paper reports PUGeo-Net as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Neural Points: Point Cloud Representation With Neural Fields for Arbitrary Upsampling (CVPR 2022 · better on 3 of 4 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 0.90)
- Arbitrary-Scale Point Cloud Upsampling by Voxel-Based Network with Latent Geometric-Consistent Learning (AAAI 2024 · similarity 0.90)
- GeoUDF: Surface Reconstruction from 3D Point Clouds via Geometry-guided Distance Representation (ICCV 2023 · similarity 0.90)
- Flattening-Net: Deep Regular 2D Representation for 3D Point Cloud Analysis (TPAMI 2023 · similarity 0.89)
- SPU-PMD: Self-Supervised Point Cloud Upsampling via Progressive Mesh Deformation (CVPR 2024 · similarity 0.89)
- Learning Continuous Implicit Field with Local Distance Indicator for Arbitrary-Scale Point Cloud Upsampling (AAAI 2024 · similarity 0.88)
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.
