频率感知流匹配用于高质量图像生成
Frequency-Aware Flow Matching for High-Quality Image Generation
把频率分解与时间相关的自适应加权引入流匹配的潜空间生成，为提升扩散/流模型的细节保真度提供了一个可插拔的条件注入思路，可直接迁移到视频生成或高分辨率合成中。
原文依据 we propose Frequency-Aware Flow Matching (FreqFlow), a novel approach that explicitly incorporates frequency-aware conditioning into the flow matching framework
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论流匹配模型通过学习逆转一个逐步加入高斯噪声的破坏过程，已成为实现逼真图像生成的强大框架。然而，由于噪声是在潜变量域中注入的，其对不同频率成分的影响并不均匀。因此，在推理阶段，流匹配模型倾向于在早期生成低频成分（全局结构），而高频成分（精细细节）只在逆向过程的后期才出现。基于这一观察，我们提出频率感知流匹配（FreqFlow），一种通过时间相关的自适应加权将频率感知条件显式融入流匹配框架的新方法。我们引入双分支架构：（1）频率分支，分别处理低频与高频成分，以捕捉全局结构并细化纹理与边缘；（2）空间分支，在潜变量域中合成图像，并由频率分支的输出进行引导。通过将频率信息显式整合进生成过程，FreqFlow 确保大尺度一致性与细粒度细节都得到有效建模——低频条件强化全局结构，高频条件提升纹理保真度与细节锐度。在类别条件 ImageNet-256 生成基准上，本方法以 1.38 的弗雷歇初始距离取得最优性能，分别比先前的扩散模型 DiT 和流匹配模型 SiT 低 0.79 和 0.58。代码见 https://github.com/OliverRensu/FreqFlow。
英文原文摘要 1645 字符
Flow matching models have emerged as a powerful framework for realistic image generation by learning to reverse a corruption process that progressively adds Gaussian noise. However, because noise is injected in the latent domain, its impact on different frequency components is non-uniform. As a result, during inference, flow matching models tend to generate low-frequency components (global structure) in the early stages, while high-frequency components (fine details) emerge only later in the reverse process. Building on this insight, we propose Frequency-Aware Flow Matching (FreqFlow), a novel approach that explicitly incorporates frequency-aware conditioning into the flow matching framework via time-dependent adaptive weighting. We introduce a two-branch architecture: (1) a frequency branch that separately processes low- and high-frequency components to capture global structure and refine textures and edges, and (2) a spatial branch that synthesizes images in the latent domain, guided by the frequency branch's output. By explicitly integrating frequency information into the generation process, FreqFlow ensures that both large-scale coherence and fine-grained details are effectively modeled--low-frequency conditioning reinforces global structure, while high-frequency conditioning enhances texture fidelity and detail sharpness. On the class-conditional ImageNet-256 generation benchmark, our method achieves state-of-the-art performance with an FID of 1.38, surpassing the prior diffusion model DiT and flow matching model SiT by 0.79 and 0.58 FID, respectively. Code is available at https://github.com/OliverRensu/FreqFlow.