New work titled “S-VGGT: Structure-Aware Subscene Decomposition for Scalable 3D Foundation Models” is now available.