UniVerse：面向无分割、解耦多概念个性化的统一调制框架
UniVerse: A Unified Modulation Framework for Segmentation-Free, Disentangled Multi-Concept Personalization
在扩散 Transformer 中通过调制实现无需分割掩码的多概念解耦与组合，可直接复用于个性化生成与概念编辑，避免依赖分割监督。
原文依据 segmentation-free, disentangled multi-concept personalization in diffusion transformers
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论个性化视觉理解已取得显著进展，但现有方法在输入图像包含多个物体时，难以定位并提取特定概念。许多先前方法严重依赖基于分割的监督，或表现出较差的组合泛化能力，限制了其准确解耦与操控单个概念的能力。在本工作中，我们提出 UniVerse，一个面向扩散 Transformer 的、无分割且解耦的多概念个性化统一调制框架。我们的方法支持可组合与可分解的概念提取，无需显式分割掩码即可实现目标物体的细粒度定位与表示。UniVerse 学习将复杂场景分解为概念特定的表示，再以统一方式将其组合，从而在多样化的视觉上下文中实现稳健的个性化。通过在多个基准上的大量实验，我们证明 UniVerse 在定位精度与视觉保真度上均显著优于最优基线。定性与定量结果表明，我们的方法能够在杂乱场景中精确提取目标概念，为更灵活、可解释且个性化的视觉生成与理解铺平道路。
英文原文摘要 1313 字符
Personalized visual understanding has advanced significantly, yet existing approaches struggle to localize and extract specific concepts when input images contain multiple objects. Many prior methods rely heavily on segmentation-based supervision or exhibit poor compositional generalization, limiting their ability to accurately disentangle and manipulate individual concepts. In this work, we propose UniVerse, a Unified Modulation Framework for segmentation-free, disentangled multi-concept personalization in diffusion transformers. Our method allows for composable and decomposable concept extraction, enabling fine-grained localization and representation of target objects without explicit segmentation masks. UniVerse learns to decompose complex scenes into concept-specific representations and then compose them in a unified manner, enabling robust personalization across diverse visual contexts. Through extensive experiments on multiple benchmarks, we demonstrate that UniVerse significantly outperforms state-of-the-art baselines in both localization accuracy and visual fidelity. Qualitative and quantitative results show that our approach can precisely extract target concepts in cluttered scenes, paving the way for more flexible, interpretable, and personalized visual generation and understanding.