Abstract: Recent advancements in 3D vision foundation models have significantly enhanced the generalization ability of 3D representation learning. However, existing models primarily rely on ...