时间感知单步扩散网络用于真实世界图像超分辨率
Time-Aware One Step Diffusion Network for Real-World Image Super-Resolution
提出时间感知的VAE编码器和时间感知VSD损失，使单步扩散超分能根据时间步动态调整潜变量特征，从而更充分利用预训练SD在不同时间步的生成先验，并实现保真度与真实感的可控权衡。
原文依据 we propose a Time-Aware one-step Diffusion Network for Real-ISR (TADSR)
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论基于扩散模型的真实世界图像超分辨率（Real-ISR）方法已展现出令人印象深刻的性能。为实现高效的Real-ISR，许多工作采用变分分数蒸馏（VSD）来蒸馏预训练的稳定扩散（SD）模型，以固定时间步进行单步超分辨率。然而，由于SD在不同时间步执行不同的生成先验，固定时间步难以使这些方法充分利用SD中的生成先验，导致性能次优。为解决此问题，我们提出了一种用于Real-ISR的时间感知单步扩散网络（TADSR）。我们首先引入时间感知变分自编码器编码器，它根据时间步将同一图像投影为不同的潜变量特征。通过时间步与潜变量特征的联合动态变化，学生模型能更好地与预训练SD的输入模式分布对齐，从而更有效地利用SD的生成能力。为更好地激活SD在不同时间步的生成先验，我们提出了一种时间感知VSD损失，它桥接学生模型与教师模型的时间步，从而产生以时间步为条件的更一致的生成先验指导。此外，尽管在不同时间步利用了SD中的生成先验，我们的方法可以通过改变时间步自然地实现保真度与真实感之间的可控权衡。实验结果表明，我们的方法仅需单步即可实现最优性能和可控的超分辨率结果。源代码已发布于https://github.com/zty557/TADSR。
英文原文摘要 1627 字符
Diffusion-based real-world image super-resolution (Real-ISR) methods have demonstrated impressive performance. To achieve efficient Real-ISR, many works employ Variational Score Distillation (VSD) to distill a pre-trained stable-diffusion (SD) model for one-step SR with a fixed timestep. However, since SD performs different generative priors at different timesteps, a fixed timestep is difficult for these methods to fully leverage the generative priors in SD, leading to suboptimal performance. To address this issue, we propose a Time-Aware one-step Diffusion Network for Real-ISR (TADSR). We first introduce a Time-Aware VAE Encoder, which projects the same image into different latent features based on timesteps. Through joint dynamic variation of timesteps and latent features, the student model can better align with the input pattern distribution of the pre-trained SD, thereby enabling more effective utilization of SD's generative capabilities. To better activate the generative prior of SD at different timesteps, we propose a Time-Aware VSD loss that bridges the timesteps of the student model and those of the teacher model, thereby producing more consistent generative prior guidance conditioned on timesteps. Additionally, though utilizing the generative prior in SD at different timesteps, our method can naturally achieve controllable trade-offs between fidelity and realism by changing the timestep. Experimental results demonstrate that our method achieves both state-of-the-art performance and controllable SR results with only a single step. The source code is released at https://github.com/zty557/TADSR