1 paper
Jiali Yao, Xinran Deng, Xin Gu +6
In this paper, we propose spatio-temporal omni-object video grounding, dubbed OmniSTVG, a new STVG task that aims at localizing spatially and temporally all targets mentioned in th…