arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.26503cs.CV
收藏

Delta Score Matters! 空间自适应多引导在扩散模型中

Delta Score Matters! Spatial Adaptive Multi Guidance in Diffusion Models

The Hong Kong University of Science and Technology (Guangzhou)(香港理工大学(广州));Griffith University & Data61/CSIRO(格里菲斯大学及Data61/CSIRO)

Haosen Li, Wenshuo Chen, Lei Wang, Shaofeng Liang, Bowen Tian, Soning Lai, Yutao Yue

AI总结:

本文提出空间自适应多引导方法,通过几何分析解决扩散模型中细节与伪影的矛盾,提升语义对齐和结构完整性。

摘要:

扩散模型在合成复杂静态和时间视觉方面取得了显著成功,这一突破主要由分类器自由引导(CFG)驱动。然而,尽管CFG在对齐生成内容与文本提示方面起关键作用,标准CFG依赖于全局均匀标量。这种同质放大导致模型陷入已记录的“细节-伪影困境”:低引导尺度无法注入复杂语义,而高尺度不可避免地导致结构退化、颜色过饱和和视频时间不一致。本文通过微分几何的视角揭示了这一缺陷的根本原因。通过分析Tweedie公式,我们揭示CFG本质上执行的是切线线性外推。由于自然数据流形高度弯曲,这种均匀线性步骤引入了严重的正交偏差。为了保持生成轨迹安全受限,我们为空间和自适应引导制定了理论上限。基于这些几何见解,我们提出了空间自适应多引导(SAMG),一种无需训练且几乎零成本的采样算法。SAMG动态计算点条件引导能量,在高能边界区域应用保守的最小尺度以保持精细微纹理,在低能区域部署激进的最大尺度以最大化语义注入。在多样化的图像(SD 1.5,SDXL,SD3.5 Medium)和视频(CogVideoX,ModelScope)架构上进行的广泛实验表明,SAMG有效解决了细节-伪影困境,实现了优越的语义对齐、结构完整性和时间平滑性,而无需任何计算开销。

英文摘要:

Diffusion models have achieved remarkable success in synthesizing complex static and temporal visuals, a breakthrough largely driven by Classifier-Free Guidance (CFG). However, despite its pivotal role in aligning generated content with textual prompts, standard CFG relies on a globally uniform scalar. This homogeneous amplification traps models in a well-documented "detail-artifact dilemma": low guidance scales fail to inject intricate semantics, while high scales inevitably cause structural degradation, color over-saturation, and temporal inconsistencies in videos. In this paper, we expose the physical root of this flaw through the lens of differential geometry. By analyzing Tweedie's Formula, we reveal that CFG intrinsically performs a tangential linear extrapolation. Because the natural data manifold is highly curved, this uniform linear step introduces a severe orthogonal deviation. To keep the generation trajectory safely bounded, we formulate a theoretical upper bound for spatial and adaptive guidance. Based on these geometric insights, we propose Spatial Adaptive Multi Guidance (SAMG), a training-free and virtually zero-cost sampling algorithm. SAMG dynamically computes point-wise conditional guidance energy, applying a conservative minimum scale to high-energy boundary regions to preserve delicate micro-textures, while deploying an aggressive maximum scale in low-energy regions to maximize semantic injection. Extensive experiments across diverse image (SD 1.5, SDXL, SD3.5 Medium) and video (CogVideoX, ModelScope) architectures demonstrate that SAMG effectively resolves the detail-artifact dilemma, achieving superior semantic alignment, structural integrity, and temporal smoothness without any computational overhead.

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。

↑