通过修正噪声水平提升扩散模型生成质量
Enhancing Sample Generation of Diffusion Models using Noise Level Correction
- 用预训练去噪网络构建噪声校正模块，动态优化生成过程中的噪声估计
- 在图像修复任务中显著提升生成质量，尤其在补全、去模糊等场景下效果明显
- 兼容现有调度器，适合需要高质量图像生成的研究与应用
扩散模型的去噪过程可视为将含噪样本近似投影到数据流形上。样本中的噪声水平近似反映其与真实数据流形的距离。基于此，我们提出一种新方法：在去噪过程中，通过校正噪声估计值，使其更贴近样本与流形的真实距离。具体地，引入一个噪声水平校正网络，利用预训练的去噪网络来精细化噪声估计。此外，我们将该方法扩展至多种图像修复任务，包括图像补全、去模糊、超分辨率、着色和压缩感知，并引入任务特定约束。实验表明，该方法在无约束和有约束生成场景下均显著提升样本质量。值得注意的是，所提出的噪声水平校正框架可与现有去噪调度器（如DDIM）兼容，并带来额外性能增益。
原文摘要 · Abstract (English)
The denoising process of diffusion models can be interpreted as an approximate projection of noisy samples onto the data manifold. Moreover, the noise level in these samples approximates their distance to the underlying manifold. Building on this insight, we propose a novel method to enhance sample generation by aligning the estimated noise level with the true distance of noisy samples to the manifold. Specifically, we introduce a noise level correction network, leveraging a pre-trained denoising network, to refine noise level estimates during the denoising process. Additionally, we extend this approach to various image restoration tasks by integrating task-specific constraints, including inpainting, deblurring, super-resolution, colorization, and compressed sensing. Experimental results demonstrate that our method significantly improves sample quality in both unconstrained and constrained generation scenarios. Notably, the proposed noise level correction framework is compatible with existing denoising schedulers (e.g., DDIM), offering additional performance improvements.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。