1 paper
Chaeyun Kim, Seunghoon Yi, Yejin Kim +2
Referring Image Segmentation (RIS) requires identifying objects from images based on textual descriptions. We observe that existing methods significantly underperform on motion-rel…