1 paper
Feng Zhu, Zongxin Yang, Xin Yu +2
Modeling temporal information for both detection and tracking in a unified framework has been proved a promising solution to video instance segmentation (VIS). However, how to effe…