Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

AdLift: Lifting Adversarial Perturbations to Safeguard 3D Gaussian Splatting Assets Against Instruction-Driven Editing

About

Recent studies have extended diffusion-based instruction-driven 2D image editing pipelines to 3D Gaussian Splatting (3DGS), enabling faithful manipulation of 3DGS assets and greatly advancing 3DGS content creation. However, it also exposes these assets to serious risks of unauthorized editing and malicious tampering. Although imperceptible adversarial perturbations against diffusion models have proven effective for protecting 2D images, applying them to 3DGS encounters two major challenges: view-generalizable protection and balancing invisibility with protection capability. In this work, we propose the first editing safeguard for 3DGS, termed AdLift, which prevents instruction-driven editing across arbitrary views and dimensions by lifting strictly bounded 2D adversarial perturbations into 3D Gaussian-represented safeguard. To ensure both adversarial perturbations effectiveness and invisibility, these safeguard Gaussians are progressively optimized across training views using a tailored Lifted PGD, which first conducts gradient truncation during back-propagation from the editing model at the rendered image and applies projected gradients to strictly constrain the image-level perturbation. Then, the resulting perturbation is backpropagated to the safeguard Gaussian parameters via an image-to-Gaussian fitting operation. We alternate between gradient truncation and image-to-Gaussian fitting, yielding consistent adversarial-based protection performance across different viewpoints and generalizes to novel views. Empirically, qualitative and quantitative results demonstrate that AdLift effectively protects against state-of-the-art instruction-driven 2D image and 3DGS editing.

Ziming Hong, Tianyu Huang, Runnan Chen, Shanshan Ye, Mingming Gong, Bo Han, Tongliang Liu• 2025

Related benchmarks

TaskDatasetResultRank
Instruction-based Image EditingInstruction-based image editing dataset Novel Views averaged across four scenes
SSIM0.7182
7
Instruction-based Image EditingInstruction-based image editing dataset averaged across four scenes (train)
SSIM81.33
7
Instruction-based Global 3DGS Editing3DGS Scenes (train Views)
CLIP Directionality13.73
6
Instruction-based Global 3DGS Editing3DGS Scenes Novel Views
CLIPd0.118
6
Instruction-based Local 3DGS Editing3DGS Scenes (train)
CLIPd0.0389
2
Instruction-based Local 3DGS Editing3DGS Scenes Novel Views
CLIPd0.0425
2
Showing 6 of 6 rows

Other info

Follow for update