MeViS: A Multi-Modal Dataset for Referring Motion Expression Video Segmentation Paper • 2512.10945 • Published Dec 11, 2025 • 1
