1 paper
Seokju Cho, Jiahui Huang, Jisu Nam +3
We introduce LocoTrack, a highly accurate and efficient model designed for the task of tracking any point (TAP) across video sequences. Previous approaches in this task often rely…