DreamStyle3D: Efficient 3D Stylized Asset Generation via Dual-Attention Disentanglement

2026-07-27

AI Summary

Existing approaches to generating stylized 3D assets often struggle to simultaneously preserve style fidelity, geometric consistency, and computational efficiency, typically relying on indirect 2D-to-3D transfer pipelines. This work proposes DreamStyle3D, a novel framework that, for the first time, explicitly disentangles geometry and style representations directly in native 3D space. By introducing a decoupled dual cross-attention mechanism, the method enables efficient style injection while maintaining structural coherence. Leveraging a lightweight training strategy and a large-scale, automatically curated dataset of approximately 15,000 geometry-style-reference triplets, DreamStyle3D supports end-to-end generation of high-fidelity, geometrically consistent stylized 3D assets in under 10 seconds, significantly outperforming current state-of-the-art methods.

Abstract

With the growth of gaming, animation, and virtual reality industries, the demand for efficient generation of stylized 3D assets is rapidly increasing. However, existing approaches still struggle to jointly preserve style fidelity, geometric consistency, and generation efficiency, as most of them still rely on indirect 2D-to-3D stylization pipelines. This motivates a native 3D stylization framework that can explicitly disentangle style from geometry while remaining efficient. To this end, we propose DreamStyle3D, an efficient framework for stylized 3D asset generation built on a Decoupled Dual Cross-Attention mechanism. Our method explicitly separates geometric and stylistic features to enable efficient style injection while preserving structural consistency, and further adopts a lightweight training strategy to enhance style consistency and model generalization. In addition, we build an automated data pipeline and construct a dataset of about 15K content-style-stylized triplets for training and evaluation. Extensive experiments demonstrate that our DreamStyle3D can generate high-fidelity, geometrically consistent stylized 3D assets within 10 seconds, substantially improving efficiency while maintaining superior style quality and offering a new solution for 3D content creation. The code and data are available at https://github.com/HVision-NKU/DreamStyle3D.

Problem

• 3D stylization
• style-geometry disentanglement
• efficient 3D generation
• stylized 3D assets
• 3D content creation

Innovation

• 3D stylization
• dual cross-attention
• feature disentanglement
• efficient generation
• geometric consistency