Return
DDAN: Dual-Dimensional Attention Network for image–text retrieval
R
D
X
DOI:10.1016/j.patcog.2026.114556.png)
Abstract
En 中文
• A dynamic element-wise attention scheme is given to enhance feature representation. • DDAN incorporates horizontal and vertical insights to enhance weights and features. • DSM is proposed to create diverse features, enriching instance representation.
Keywords:
Image–text retrieval
Attention scheme
Polysemous embedding
Cross-modal matching
Journal
IF:
7.6
Papers:
1.3W
Citations:
4.5W
Organization
No organization information available
