arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2512.09477cs.CVcs.AI
收藏

扩散模型潜在空间中的颜色编码

Color encoding in Latent Space of Stable Diffusion Models

Computer Vision Center / Universitat Autònoma de Barcelona(计算机视觉中心 / 巴塞罗那自治大学)

Guillem Arias, Ariadna Solà, Martí Armengod, Maria Vanrell

AI总结:

本文研究了Stable Diffusion模型中颜色在潜在空间中的编码机制,揭示了颜色、亮度和形状在不同潜在通道中的表示方式,为生成模型的理解与改进提供了新视角。

摘要:

最近扩散生成模型的进展在视觉保真度上取得了显著成就,但对特定感知属性如颜色和形状如何内部表示仍缺乏深入理解。本文通过系统分析Stable Diffusion的潜在表示,探讨了颜色在生成模型中的编码方式。通过受控合成数据集、主成分分析(PCA)和相似性度量,我们发现颜色信息主要沿环形、对立轴编码,这些轴主要捕捉在潜在通道c_3和c_4中,而亮度和形状主要表示在通道c_1和c_2中。我们的发现表明,Stable Diffusion的潜在空间表现出与高效编码表示一致的可解释结构。这些见解为未来模型理解、编辑应用以及更解耦的生成框架设计提供了基础。

英文摘要:

Recent advances in diffusion-based generative models have achieved remarkable visual fidelity, yet a detailed understanding of how specific perceptual attributes - such as color and shape - are internally represented remains limited. This work explores how color is encoded in a generative model through a systematic analysis of the latent representations in Stable Diffusion. Through controlled synthetic datasets, principal component analysis (PCA) and similarity metrics, we reveal that color information is encoded along circular, opponent axes predominantly captured in latent channels c_3 and c_4, whereas intensity and shape are primarily represented in channels c_1 and c_2. Our findings indicate that the latent space of Stable Diffusion exhibits an interpretable structure aligned with a efficient coding representation. These insights provide a foundation for future work in model understanding, editing applications, and the design of more disentangled generative frameworks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑