Enhancing Video-Language Representations With Structural Spatio-Temporal Alignment2024-12-010 OA AI DOI:10.1109/tpami.2024.3393452原文链接原文求助分享收藏摘要 En