.png){kind=icon}
.png){kind=icon}
Return
LDAC-Net: A parallel dual-encoder network with Cross-Gated Residual Fusion for pavement crack segmentation
.png){kind=icon}
PRE
AI
G
C
Z
Y
DOI：10.1016/j.imavis.2026.106078.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Parallel dual-encoder CNN-Transformer network is proposed for crack segmentation. • Dynamic Kernel Attention improves orientation-aware responses to crack details. • Adaptive Routing Enhanced Transformer strengthens multi-scale context aggregation. • Cross-Gated Residual Fusion aligns multi-level features via gated residual weighting. • Results on three public datasets show improved accuracy and structural continuity.
Journal
IF:
4.2
Papers:
4.0K
Citations:
6.7K
Organization
No organization information available
Cited Papers