1 paper
Tengfei Liu, Yang Shi, Yuran Wang +16
Existing video captioning models generate natural descriptions of video content but cannot explicitly ground local visual elements to multiple reference images. We introduce multi-…