基于 SVD 特征分解的免训练风格个性化
A Training-Free Style-Personalization via SVD-Based Feature Decomposition
用 SVD 分解内部特征的主成分来分离风格与内容，并在推理阶段通过特征混合与注意力校正实现风格迁移，为免训练个性化生成提供了一条可复用的「特征分解+注意力干预」路径。
原文依据 the dominant singular values of the internal feature encode style-related components
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论我们提出了一个免训练的风格个性化图像生成框架，它在推理阶段基于尺度级自回归模型运行。我们的方法在单张参考风格图像的引导下生成风格化图像，同时保持语义一致性并缓解内容泄漏。通过对生成过程进行细致的逐步分析，我们识别出一个关键步骤，在该步骤中内部特征的主导奇异值编码了风格相关成分。基于这一发现，我们引入了两个轻量级控制模块：主特征混合，通过基于 SVD 的特征重建实现对风格的精确调制；结构注意力校正，在精细阶段利用内容引导的注意力校正来稳定结构一致性。在无需任何额外训练的情况下，大量实验表明，与微调基线相比，我们的方法在风格保真度和提示词保真度上取得了有竞争力的表现，同时提供更快的推理速度和更大的部署灵活性。
英文原文摘要 1049 字符
We present a training-free framework for style-personalized image generation that operates during inference using a scale-wise autoregressive model. Our method generates a stylized image guided by a single reference style while preserving semantic consistency and mitigating content leakage. Through a detailed step-wise analysis of the generation process, we identify a pivotal step where the dominant singular values of the internal feature encode style-related components. Building upon this insight, we introduce two lightweight control modules: Principal Feature Blending, which enables precise modulation of style through SVD-based feature reconstruction, and Structural Attention Correction, which stabilizes structural consistency by leveraging content-guided attention correction across fine stages. Without any additional training, extensive experiments demonstrate that our method achieves competitive style fidelity and prompt fidelity compared to fine-tuned baselines, while offering faster inference and greater deployment flexibility.