DisCo3D: Distilling Multi-View Consistency for 3D Scene Editing

TL;DR

DisCo3D introduces a framework that distills 3D multi-view consistency into a 2D editing process, enabling more stable and high-quality 3D scene editing by combining multi-view adaptation and consistency constraints.

Contribution

The paper presents DisCo3D, a novel method that distills 3D consistency priors into a 2D editor, improving multi-view consistency and editing quality in 3D scene editing.

Findings

DisCo3D achieves superior multi-view consistency.

It outperforms state-of-the-art methods in editing quality.

The framework enables stable and efficient 3D scene editing.

Abstract

While diffusion models have demonstrated remarkable progress in 2D image generation and editing, extending these capabilities to 3D editing remains challenging, particularly in maintaining multi-view consistency. Classical approaches typically update 3D representations through iterative refinement based on a single editing view. However, these methods often suffer from slow convergence and blurry artifacts caused by cross-view inconsistencies. Recent methods improve efficiency by propagating 2D editing attention features, yet still exhibit fine-grained inconsistencies and failure modes in complex scenes due to insufficient constraints. To address this, we propose \textbf{DisCo3D}, a novel framework that distills 3D consistency priors into a 2D editor. Our method first fine-tunes a 3D generator using multi-view inputs for scene adaptation, then trains a 2D editor through consistency distillation. The edited multi-view outputs are finally optimized into 3D representations via Gaussian Splatting. Experimental results show DisCo3D achieves stable multi-view consistency and outperforms state-of-the-art methods in editing quality.