Cross-Modal Attention With Semantic Consistence for Image–Text Matching | IEEE Journals & Magazine | IEEE Xplore