Abstract
Recent advancements in diffusion-based generative priors have enabled visually plausible image compression at extremely low bit rates. However, existing approaches suffer from slow sampling processes and suboptimal bit allocation due to fragmented training paradigms. In this work, we propose Accelerate \textbf{Diff}usion-based Image Compression via \textbf{C}onsistency Prior \textbf{R}efinement (DiffCR), a novel compression framework for efficient and high-fidelity image reconstruction. At the heart of DiffCR is a Frequency-aware Skip Estimation (FaSE) module that refines the $\epsilon$-prediction prior from a pre-trained latent diffusion model and aligns it with compressed latents at different timesteps via Frequency Decoupling Attention (FDA). Furthermore, a lightweight consistency estimator enables fast \textbf{two-step decoding} by preserving the semantic trajectory of diffusion sampling. Without updating the backbone diffusion model, DiffCR achieves substantial bitrate savings (27.2\% BD-rate (LPIPS) and 65.1\% BD-rate (PSNR)) and over $10\times$ speed-up compared to SOTA diffusion-based compression baselines.
Abstract (translated)
最近基于扩散模型的生成先验技术的进步已经使得在极低比特率下实现视觉上合理的图像压缩成为可能。然而，现有的方法由于碎片化的训练范式而面临着采样过程缓慢和比特分配不理想的问题。在这项工作中，我们提出了一种名为“通过一致性优先级精炼加速扩散模型图像压缩”的新框架（DiffCR），旨在高效且高保真地重建图像。DiffCR的核心是一个频率感知跳跃估计模块（FaSE），该模块能够优化预训练的潜在扩散模型中的$\epsilon$预测先验，并通过频率解耦注意力机制将其与不同时间步长下的压缩潜码对齐。此外，一个轻量级的一致性评估器能够在保持扩散采样语义轨迹的同时实现快速两阶段解码过程。 在不更新基础扩散模型的情况下，DiffCR相比最先进的基于扩散的压缩基线方法，在比特率节省方面取得了显著的进步（LPIPS指标下节约27.2%比特率，PSNR指标下节约65.1%比特率），并且速度提高了超过十倍。
URL
https://arxiv.org/abs/2601.10373
