Abstract
All-in-one image restoration aims to recover clear images from various degradation types and levels with a unified model. Nonetheless, the significant variations among degradation types present challenges for training a universal model, often resulting in task interference, where the gradient update directions of different tasks may diverge due to shared parameters. To address this issue, motivated by the routing strategy, we propose DFPIR, a novel all-in-one image restorer that introduces Degradation-aware Feature Perturbations(DFP) to adjust the feature space to align with the unified parameter space. In this paper, the feature perturbations primarily include channel-wise perturbations and attention-wise perturbations. Specifically, channel-wise perturbations are implemented by shuffling the channels in high-dimensional space guided by degradation types, while attention-wise perturbations are achieved through selective masking in the attention space. To achieve these goals, we propose a Degradation-Guided Perturbation Block (DGPB) to implement these two functions, positioned between the encoding and decoding stages of the encoder-decoder architecture. Extensive experimental results demonstrate that DFPIR achieves state-of-the-art performance on several all-in-one image restoration tasks including image denoising, image dehazing, image deraining, motion deblurring, and low-light image enhancement. Our codes are available at this https URL.
Abstract (translated)
全图像复原旨在利用统一的模型从各种退化类型和程度中恢复清晰图像。然而，不同退化类型的显著差异为训练通用模型带来了挑战，通常会导致任务干扰问题——由于共享参数，不同任务的梯度更新方向可能会发生分歧。为了应对这一问题，受路由策略启发，我们提出了DFPIR（Degradation-aware Feature Perturbations for Image Restoration），这是一种新的全图像复原方法，通过引入退化感知特征扰动（DFP）来调整特征空间以适应统一参数空间。 在本文中，特征扰动主要包括通道级扰动和注意力级扰动。具体而言，通道级扰动是通过根据不同的退化类型引导高维空间中的通道洗牌实现的；而注意力级扰动则是通过对注意力空间进行选择性屏蔽来完成的。为了达成这些目标，我们设计了一种退化导向的扰动模块（DGPB），用于实现在编码器-解码器架构的编码和解码阶段之间的这两种功能。 广泛的实验结果表明，DFPIR在包括图像去噪、图像去雾、图像除雨、运动模糊恢复以及低光照图像增强在内的几个全图像复原任务中达到了最先进的性能。我们的代码可在提供的链接地址获取。
URL
https://arxiv.org/abs/2505.12630
