Rectified Point Flow: Generic Point Cloud Pose Estimation
Tao SunStanford UniversityLiyuan ZhuStanford UniversityShengyu HuangNVIDIA ResearchShuran SongStanford UniversityIro ArmeniStanford University · 2025
Read Rectified Point Flow: Generic Point Cloud Pose Estimation by Tao SunStanford UniversityLiyuan ZhuStanford UniversityShengyu HuangNVIDIA ResearchShuran SongStanford UniversityIro ArmeniStanford University (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-05282.txt.
Description
Lining up two 3D scans of an object, or assembling scanned parts, usually takes separate tools with hand-built symmetry rules. Tao Sun and co-authors recast both jobs as one learning problem. Their method learns a flow that carries each noisy point toward its place in the finished shape. Part poses are read off the flow's end state. Symmetry is learned from data, not labeled by hand. An encoder that favors overlap regions sharpens the input.
Across six benchmarks the method set new best scores for both registration and assembly. Training jointly on many datasets let shared shape knowledge lift accuracy. A single tool for both jobs simplifies robot and scanning pipelines. The team's project page shows the flow in action.
Catalog record
| CREATOR | Tao SunStanford UniversityLiyuan ZhuStanford UniversityShengyu HuangNVIDIA ResearchShuran SongStanford UniversityIro ArmeniStanford University |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-05282