GaussianEditor: Editing 3D Gaussians Delicately with Text Instructions

Editing 3D Gaussians Delicately with
Text Instructions

CVPR 2024
Huawei Inc.
* denotes equal contributions.


Recently, impressive results have been achieved in 3D scene editing with text instructions based on a 2D diffusion model. However, current diffusion models primarily generate images by predicting noise in the latent space, and the editing is usually applied to the whole image, which makes it challenging to perform delicate, especially localized, editing for 3D scenes. Inspired by recent 3D Gaussian splatting, we propose a systematic framework, named GaussianEditor, to edit 3D scenes delicately via 3D Gaussians with text instructions. Benefiting from the explicit property of 3D Gaussians, we design a series of techniques to achieve delicate editing. Specifically, we first extract the region of interest (RoI) corresponding to the text instruction, aligning it to 3D Gaussians. The Gaussian RoI is further used to control the editing process. Our framework can achieve more delicate and precise editing of 3D scenes than previous methods while enjoying much faster training speed, i.e. within 20 minutes on a single V100 GPU, more than twice as fast as Instruct-NeRF2NeRF (45 minutes -- 2 hours).


360° Scene Editing

Multiple-Round Editing

Comparisons with Instruct-NeRF2NeRF

Complex Multi-Object Scenes

More Examples

Extension with GaussianDreamer


  author = {Fang, Jiemin and Wang, Junjie and Zhang, Xiaopeng and Xie, Lingxi and Tian, Qi},
  title = {GaussianEditor: Editing 3D Gaussians Delicately with Text Instructions},
  year = {2024},