Most existing image-text matching methods adopt triplet loss as the
opti...
Most existing cross-modal retrieval methods employ two-stream encoders w...
In the absence of an authoritative statement about a rumor, people may e...
Video captioning has been attracting broad research attention in multime...
share