Fetching the paper…

UVOSAM: A Mask-free Paradigm for Unsupervised Video Object Segmentation via Segment Anything Model · Around