R3PM-Net: Real-time, Robust, Real-world Point Matching Network
Yasaman Kashefbahrami, Erkut Akdag, Panagiotis Meletis, Evgeniya Balmashnova, Dip Goswami, Egor Bondarau · 2026
Read R3PM-Net: Real-time, Robust, Real-world Point Matching Network by Yasaman Kashefbahrami, Erkut Akdag, Panagiotis Meletis, Evgeniya Balmashnova, Dip Goswami, Egor Bondarau (2026) 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-2604-05060.txt.
Description
Kashefbahrami and five co-authors introduce R3PM-Net, a lightweight, global-aware point matching network for point cloud registration. Registration estimates a rigid transformation between two point clouds, a key task in 3D data processing. Deep methods handle noise, outliers, occlusion, and initialization, yet they train on clean, dense, synthetic data, limiting real use. R3PM-Net prioritizes generalizability and real-time efficiency to bridge that gap. The authors propose Sioux-Cranfield and Sioux-Scans as an evaluation ground for real registration.
Catalog record
| CREATOR | Yasaman Kashefbahrami, Erkut Akdag, Panagiotis Meletis, Evgeniya Balmashnova, Dip Goswami, Egor Bondarau |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50180 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.05060