Tune-Your-Style: Intensity-tunable 3D Style Transfer with Gaussian Splatting

TL;DR

This work tackles the fixed-output limitation of contemporary 3D style transfer by introducing Tune-Your-Style, an intensity-tunable framework that explicitly models style strength via Gaussian neurons and a learnable style tuner. It combines Intensity-tunable Style Injection with Tunable Stylization Guidance, leveraging cross-view diffusion priors and a two-stage optimization to achieve stable, multi-view-consistent stylization. The approach delivers sharper style textures with better fidelity to reference styles and allows users to flexibly adjust style intensity, including multi-style combinations, while maintaining content integrity. By integrating 3D Gaussian Splatting with diffusion-based guidance, the method enables practical, interactive 3D style transfer suitable for artistic creation, gaming, and visualization.

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 \textbf{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.

Figures (9)

• 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.
• 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 adjust the style intensity injected into the scene. To facilitate the learning of the style intensity and tuner, TSG first employs a diffusion model to perform style transfer on rendered views, and obtains multi-view consistent stylized results through cross-view style alignment. Then, TSG adopts a two-stage optimization strategy to achieve stable and efficient tunable stylization guidance, with full-style guidance in the stylized results and zero-style guidance in the initial rendering.
• Figure 3: Results of the qualitative comparison with 3DGS-based style transfer methods. Our method exhibits clearer style textures, maintains more consistent color fidelity with the reference style, and produces fewer artifacts. Best viewed with zoom-in.
• Figure 4: Results of intensity-tunable 3D stylization. The style tuner is marked above each column of rendered results, with the style injection intensity rising in proportion to the tuner value. Users can adjust the style tuner to their preferences to obtain desired results.
• Figure 5: Results of multi-style combination. We select two different styles to inject into the interior and exterior regions of the mask in each instance, and we can independently adjust the injection intensity of each style, thus controlling the dominant style of stylized scene.