.png){kind=icon}
.png){kind=icon}
Return
SAGA-Feat: A semantic- and geometry-aware network for sparse local feature learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.131349.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Proposes a weakly supervised network for enhanced local feature learning. • Incorporates semantic–geometric attention for robust multi-scale encoding. • Enhances structural consistency via dual-domain normalization and adaptive fusion. • Utilizes deformable sampling grids for accurate feature reconstruction. • Achieves state-of-the-art or competitive performance among learning-based methods on multiple benchmarks.
Keywords:
weakly supervised learning
semantic-geometric attention
multi-scale encoding
structural consistency
deformable sampling
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W