NANO3D: A Training-Free Approach for Efficient 3D Editing Without Masks

2025-10-16

AI Summary

Existing 3D object editing methods suffer from reliance on multi-view reconstruction, low efficiency, inconsistent outputs, and difficulty preserving unedited regions. To address these issues, this paper proposes a training-free, mask-free feed-forward 3D editing framework. Methodologically, it integrates FlowEdit with TRELLIS, enabling localized editing guided solely by a single front-view image, and introduces a region-aware Voxel/Slab-Merge fusion strategy jointly optimized with multi-view consistency constraints to ensure geometric and appearance fidelity. Key contributions include: (1) the first training-free, mask-free framework achieving high-fidelity 3D editing; (2) an adaptive region merging mechanism that significantly improves structural coherence; and (3) Nano3D-Edit-100k—the first large-scale 3D editing benchmark comprising 100,000 paired samples. Extensive experiments demonstrate that our approach substantially outperforms state-of-the-art methods in both 3D consistency and visual quality.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

3D object editing is essential for interactive content creation in gaming, animation, and robotics, yet current approaches remain inefficient, inconsistent, and often fail to preserve unedited regions. Most methods rely on editing multi-view renderings followed by reconstruction, which introduces artifacts and limits practicality. To address these challenges, we propose Nano3D, a training-free framework for precise and coherent 3D object editing without masks. Nano3D integrates FlowEdit into TRELLIS to perform localized edits guided by front-view renderings, and further introduces region-aware merging strategies, Voxel/Slat-Merge, which adaptively preserve structural fidelity by ensuring consistency between edited and unedited areas. Experiments demonstrate that Nano3D achieves superior 3D consistency and visual quality compared with existing methods. Based on this framework, we construct the first large-scale 3D editing datasets Nano3D-Edit-100k, which contains over 100,000 high-quality 3D editing pairs. This work addresses long-standing challenges in both algorithm design and data availability, significantly improving the generality and reliability of 3D editing, and laying the groundwork for the development of feed-forward 3D editing models. Project Page:https://jamesyjl.github.io/Nano3D

Problem

• Efficient 3D object editing without masks for interactive content creation
• Precise localized edits while preserving structural fidelity in unedited regions
• Addressing artifacts from multi-view rendering reconstruction in current methods

Innovation

• Training-free framework for precise 3D editing
• FlowEdit integration enables localized edits without masks
• Voxel/Slat-Merge strategies preserve structural fidelity