Corpus Video Moment Retrieval

Corpus Video Moment Retrieval involves temporal localization within all videos of a dataset based on a single textual query, to identify the moments relevant to the query. This task aims to enhance the accuracy and efficiency of multimodal information retrieval and holds extensive application value, such as in video surveillance, content recommendation, and intelligent search.