Diving Into The Relations: Leveraging Semantic and Visual Structures For Video Moment Retrieval | IEEE Conference Publication | IEEE Xplore