Chat-Edit-3D++ is a dialogue-based approach for editing 3D and 4D scenes that uses an LLM to interpret flexible text input and autonomously invoke visual editing tools. Its Hash-Atlas network reformulates 3D scene editing as operations on 2D atlas images, decoupling 2D editing from 3D reconstruction, and the system extends to monocular 4D scenes through motion constraints and an LLM fine-tuned on trajectory data, coordinating multiple visual models for diverse editing effects.
