.png){kind=icon}
.png){kind=icon}
Return
Transformative attention fusion: a multi-directional feature fusion module
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2026.134617.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A dual-stage reference-guided fusion is proposed for integrating transformed-view features in camouflaged object detection. • DCA uses the original view to select compact directional branch descriptors. • PVG learns pixel-wise view weights and outperforms uniform averaging and hard selection. • Five-run experiments show strong and consistent results on COD10K and NC4K. • Corruption tests show improved stability under global and single-view degradation.
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W