AdaIR is an all-in-one image restoration model that addresses various forms of image degradation, including noise, blur, haze, and rain. The model leverages frequency mining and modulation to adaptively remove these degradations without prior knowledge of the specific type of corruption. The key contributions of AdaIR include: 1. **Frequency Mining Module (FMiM)**: Extracts low- and high-frequency features from the input image using an adaptive spectra decomposition guided by the degraded image's spectrum. 2. **Frequency Modulation Module (FMoM)**: Refines the extracted features by facilitating interactions between different frequency components, enhancing the restoration quality. 3. **Adaptive Frequency Learning Block (AFLB)**: A plugin block designed to integrate frequency mining and modulation into existing image restoration architectures. AdaIR achieves state-of-the-art performance on multiple image restoration tasks, including denoising, dehazing, deraining, motion deblurring, and low-light image enhancement. Extensive experiments demonstrate the effectiveness of AdaIR, with significant improvements over competing methods in terms of PSNR and SSIM metrics. The model's ability to adaptively handle different degradation types and its efficient computational requirements make it a promising solution for real-world image restoration tasks.AdaIR is an all-in-one image restoration model that addresses various forms of image degradation, including noise, blur, haze, and rain. The model leverages frequency mining and modulation to adaptively remove these degradations without prior knowledge of the specific type of corruption. The key contributions of AdaIR include: 1. **Frequency Mining Module (FMiM)**: Extracts low- and high-frequency features from the input image using an adaptive spectra decomposition guided by the degraded image's spectrum. 2. **Frequency Modulation Module (FMoM)**: Refines the extracted features by facilitating interactions between different frequency components, enhancing the restoration quality. 3. **Adaptive Frequency Learning Block (AFLB)**: A plugin block designed to integrate frequency mining and modulation into existing image restoration architectures. AdaIR achieves state-of-the-art performance on multiple image restoration tasks, including denoising, dehazing, deraining, motion deblurring, and low-light image enhancement. Extensive experiments demonstrate the effectiveness of AdaIR, with significant improvements over competing methods in terms of PSNR and SSIM metrics. The model's ability to adaptively handle different degradation types and its efficient computational requirements make it a promising solution for real-world image restoration tasks.