[论文解读] Tune-Your-Style: Intensity-tunable 3D Style Transfer with Gaussian Splatting

Yian Zhao, Rushi Ye+7 more|arXiv (Cornell University)|Jan 31, 2026

Generative Adversarial Networks and Image Synthesis被引用 0

一句话总结

简述：通过高斯点样化实现强度可调的三维风格迁移，建模风格强度的高斯神经元与可学习的风格调节器，以及通过扩散引导的可调风格化实现多视角一致性。

ABSTRACT

3D style transfer refers to the artistic stylization of 3D assets based on reference style images. Recently, 3DGS-based stylization methods have drawn considerable attention, primarily due to their markedly enhanced training and rendering speeds. However, a vital challenge for 3D style transfer is to strike a balance between the content and the patterns and colors of the style. Although the existing methods strive to achieve relatively balanced outcomes, the fixed-output paradigm struggles to adapt to the diverse content-style balance requirements from different users. In this work, we introduce a creative intensity-tunable 3D style transfer paradigm, dubbed extbf{Tune-Your-Style}, which allows users to flexibly adjust the style intensity injected into the scene to match their desired content-style balance, thus enhancing the customizability of 3D style transfer. To achieve this goal, we first introduce Gaussian neurons to explicitly model the style intensity and parameterize a learnable style tuner to achieve intensity-tunable style injection. To facilitate the learning of tunable stylization, we further propose the tunable stylization guidance, which obtains multi-view consistent stylized views from diffusion models through cross-view style alignment, and then employs a two-stage optimization strategy to provide stable and efficient guidance by modulating the balance between full-style guidance from the stylized views and zero-style guidance from the initial rendering. Extensive experiments demonstrate that our method not only delivers visually appealing results, but also exhibits flexible customizability for 3D style transfer. Project page is available at https://zhao-yian.github.io/TuneStyle.

研究动机与目标

• 通过实现用户可控的内容-风格平衡，解决三维风格迁移中的固定输出限制。
• 对风格强度进行显式建模，并提供在三维高斯点样场景中注入风格的可调机制。
• 通过跨视图风格对齐和基于扩散的引导实现多视角一致性。
• 提供两阶段优化策略，在稳定可调风格化的同时保持效率。

提出的方法

• 将场景表示为三维高斯原始元素，并通过三维高斯抹光（3DGS）进行渲染。
• 引入高斯神经元以预测每个高斯原始元素的属性偏移量，从而实现风格强度的预测。
• 定义一个带有阶梯函数的可调风格注入器，将连续的调节输入映射到离散嵌入。
• 应用三维高斯滤波器去除冗余原始元素并降低伪影。
• 使用二维扩散模型对渲染视图进行风格化，并通过跨视图对齐引导三维更新。
• 实施两阶段优化：先进行全风格引导，再进行带有零风格与全风格项的可调引导。
• 跨视图风格对齐注入锚视图特征并进行内容标定以保持三维一致性。

Figure 1 : (a) Existing fixed-output paradigm struggles to adapt to the diverse content-style balance requirements. (b) Our intensity-tunable 3D style transfer paradigm enables users to flexibly adjust the style intensity to achieve the desired content-style balance.

实验结果

研究问题

• RQ1 我们如何在三维风格迁移中建模并控制风格强度？
• RQ2 强度可调框架是否在内容-风格平衡方面优于固定输出方法？
• RQ3 基于扩散的先验和跨视图对齐是否能为三维场景提供稳定且多视角一致的风格化？
• RQ4 两阶段引导策略对风格化质量与可调性有何影响？

主要发现

• 所提出的方法在多视角一致性方面优于基线方法（短程与长程 LPIPS/RMSE 指标对比）。
• 基于 CLIP 的风格保真度指标（CLIP S 与 CLIP S_dir）高于基线方法。
• 用户研究显示所提出的方法更受偏好，整体观感更好。
• 强度可调性实现对内容-风格平衡的灵活控制，在定性结果与多风格组合中得到验证。
• 两阶段优化对稳定风格化和有效学习风格调节器至关重要。

Figure 2 : Overall framework. Our method comprise two pivotal components, namely Intensity-tunable Style Injection (ISI) and Tunable Stylization Guidance (TSG). ISI introduces Gaussian neurons to explicitly model style intensity and parameterizes a learnable style tuner, enabling users to flexibly a