计算机视觉

CoRef-GS: Cooperative Referring Gaussian Splatting for Multi-Agent Scene Understanding

Referring scene understanding for embodied robots requires grounding object- and relation-centric language queries from