ReferEverything: Towards Segmenting Everything We Can Speak of in Videos Paper • 2410.23287 • Published 7 days ago • 17