EditFlow3D: Automated Local Editing of 3D Assets with Trajectory Preservation

2026-08-04

AI Summary

Existing methods for local 3D editing struggle to accurately localize target regions while preserving the structural and appearance consistency of non-edited areas. This work proposes a training-free local 3D editing framework that leverages vision-language models to interpret editing intent, automatically generating guidance images and refined 3D masks, and performs edits directly in the native latent space of a pre-trained 3D generative model. The approach introduces a mask-guided diffusion flow to focus editing on the target region and incorporates a progressive trajectory preservation mechanism to maintain consistency in non-target regions. Furthermore, the authors introduce EditFlow-Bench, a more comprehensive benchmark for evaluating local 3D editing. Experiments demonstrate that the proposed method significantly improves both editing accuracy and fidelity of non-edited regions on both Edit3D-Bench and EditFlow-Bench, as validated by user studies and quantitative metrics.

Abstract

Controllable local editing of 3D assets requires precise target localization and appropriate visual guidance. However, existing methods lack a simple yet accurate way to obtain 3D masks and struggle to achieve the desired edit while faithfully preserving the structure and appearance of non-target regions. To address these challenges, we present EditFlow3D, a training-free framework for local 3D editing. Given a source asset and an edit instruction, a VLM-driven workflow interprets the editing intent and automatically constructs a visual guidance image and a refined 3D editing mask, enabling localized editing in the native representation space of a pretrained 3D generative model. Specifically, mask-guided differential flow focuses the edit on the target region, while step-wise trajectory preservation maintains consistency between non-target regions and the source asset without directly replacing intermediate features. Since the existing Edit3D-Bench covers only a limited range of local editing categories, we further introduce EditFlow-Bench as a complementary benchmark encompassing a broader variety of structural and appearance edits, and evaluate EditFlow3D on both benchmarks. Quantitative results, qualitative comparisons, and a user study demonstrate that EditFlow3D achieves more accurate target-region editing and better preserves non-target regions than existing 3D editing methods.

Problem

• local editing
• 3D assets
• trajectory preservation
• 3D mask
• appearance consistency

Innovation

• 3D local editing
• trajectory preservation
• mask-guided differential flow
• training-free framework
• visual-language model