45 citations
1 paper
Haiwen Diao, Ying Zhang, Wei Liu +2
Exploiting fine-grained correspondence and visual-semantic alignments has shown great potential in image-text matching. Generally, recent approaches first employ a cross-modal atte…