.png){kind=icon}
.png){kind=icon}
Return
DGS-Net: A dual geometric self-supervised network for 3D mesh segmentation
.png){kind=icon}
PRE
AI
J
Y
DOI：10.1016/j.eswa.2026.133919.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• DGS-Net: A self-supervised framework for robust 3D mesh segmentation. • Adaptive Gated Fusion balances extrinsic and intrinsic geometric features. • Geometric Positional Encoding enriches tokens with multi-order priors. • Dual SSL tasks combine masked reconstruction and contrastive clustering.
Keywords:
Mesh segmentation
Self-supervised learning
Geometric encoding
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W
Organization
No organization information available
Cited Papers