Abstract: As a new emerging task, video corpus moment retrieval (VCMR) aims to find the video segments relevant to a given natural language query from a large number of untrimmed videos. It mainly ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results