arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.05534cs.CV
收藏

SSG: 多尺度视觉自回归生成的缩放空间引导

SSG: Scaled Spatial Guidance for Multi-Scale Visual Autoregressive Generation

Korea Advanced Institute of Science and Technology(韩国科学技术院)

Youngwoo Shin, Jiwan Hur, Junmo Kim

AI总结:

SSG通过缩放空间引导方法,在无需训练的情况下提升多尺度视觉自回归生成的保真度和多样性,同时保持低延迟。

摘要:

视觉自回归(VAR)模型通过逐级预测生成图像,自然实现从粗到细、快速且高保真的合成,模仿人类感知。在实践中,这种层次结构在推理时可能会漂移,因为有限的容量和累积的误差导致模型偏离其从粗到细的性质。我们从信息论的角度重新审视这一限制,并推导出确保每个尺度贡献未被早期尺度解释的高频内容可以缓解训练-推理差异。基于这一见解,我们提出了缩放空间引导(SSG),一种无需训练的推理时引导方法,能够引导生成朝着预期的层次结构发展,同时保持全局一致性。SSG强调目标高频信号,定义为语义残差,从较粗的先验中分离出来。为了获得这一先验,我们利用一种原理性的频域过程,即离散空间增强(DSE),该过程旨在通过频域感知的构造来增强和更好地隔离语义残差。SSG适用于利用离散视觉标记的VAR模型,无论标记化设计或条件模式如何。实验表明,SSG在保真度和多样性方面获得了一致的提升,同时保持低延迟,揭示了在从粗到细图像生成中未被利用的效率。代码可在https://github.com/Youngwoo-git/SSG上获得。

英文摘要:

Visual autoregressive (VAR) models generate images through next-scale prediction, naturally achieving coarse-to-fine, fast, high-fidelity synthesis mirroring human perception. In practice, this hierarchy can drift at inference time, as limited capacity and accumulated error cause the model to deviate from its coarse-to-fine nature. We revisit this limitation from an information-theoretic perspective and deduce that ensuring each scale contributes high-frequency content not explained by earlier scales mitigates the train-inference discrepancy. With this insight, we propose Scaled Spatial Guidance (SSG), training-free, inference-time guidance that steers generation toward the intended hierarchy while maintaining global coherence. SSG emphasizes target high-frequency signals, defined as the semantic residual, isolated from a coarser prior. To obtain this prior, we leverage a principled frequency-domain procedure, Discrete Spatial Enhancement (DSE), which is devised to sharpen and better isolate the semantic residual through frequency-aware construction. SSG applies broadly across VAR models leveraging discrete visual tokens, regardless of tokenization design or conditioning modality. Experiments demonstrate SSG yields consistent gains in fidelity and diversity while preserving low latency, revealing untapped efficiency in coarse-to-fine image generation. Code is available at https://github.com/Youngwoo-git/SSG.

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。

↑