1 paper
Kunyu Peng, Jia Fu, Kailun Yang +8
We introduce a new task called Referring Atomic Video Action Recognition (RAVAR), aimed at identifying atomic actions of a particular person based on a textual description and the…