CLIP-GS efficiently models compact and view-consistent semantic representations using 3D Gaussians, enabling real-time and precise 3D semantic understanding.
