.png){kind=icon}
.png){kind=icon}
Return
PU-MG: Mutual guidance framework for Point Cloud Upsampling
.png){kind=icon}
PRE
AI
DOI：10.1016/j.displa.2025.103270.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• PU-MG introduces a mutual guidance framework for denoising-based upsampling. • PU-MG couples sub-networks to ease conflicts and improve convergence. • AGFFM fuses features by guiding perturbation disentanglement in latent space. • DGR loss aligns perturbed and clean points to ensure accurate reconstruction. • Experiments on four benchmarks show strong performance and robustness.
Keywords:
Point Cloud Upsampling
Deep learning
Mutual guidance