1 paper
Yunhao Li, Qin Li, Hao Wang +5
Current multi-object tracking (MOT) aims to predict trajectories of targets (i.e., ''where'') in videos. Yet, knowing merely ''where'' is insufficient in many crucial applications.…