1 paper · 1 filter
Ce Bian, Xusheng He, Jinrong Zhang +3
This report presents a two-stage, training-free solution for the MeViS-Text track of the 8th LSVOS Challenge. The task requires a model to localize and segment the object specified…