JPEG compression brings artifacts into the compressed image, which not only degrade visual quality, but also affect the performance of other image processing tasks.
To address this issue, many learning-based compression artifacts removal methods have been developed in recent years, with remarkable success. However, existing learningbased methods generally only exploit spatial information and lack exploration of frequency domain information.
Exploring frequency domain information is critical because JPEG compressionis actually performed in the frequency domain using the Discrete Cosine Transform (DCT). To effectively leverage information from both the spatial and frequency domains, we propose a novel Dual-Domain Learning Network for JPEG artifacts removal (D2LNet).
Our approach first transforms the spatial domain image to the frequency domain by the fast Fourier transform (FFT). We then introduce two core modules, Amplitude Correction Module (ACM) and Phase Correction Module (PCM), which facilitate interactive learning of spatial and frequency domain information.
The architecture of our network, which consists of two main modules:
the Amplitude Correction Module (ACM) and the Phase Correction Module (PCM).
Specifically, the ACM restores the amplitude spectrum of degraded images to remove JPEG artifacts, and the PCM restores the phase spectrum information to refine the highfrequency information.
The following is the main architecture of our Dual-Domain Learning Network:
Table 1.
PSNR/SSIM/PSNR-B results of our method comparaed to other nine methods in three datasets, with the best outcomes being highlighted in red.
Table 2.
PSNR/SSIM/PSNR-B results of different methods on the three color datasets, with the best outcomes being highlighted in red.
Table 3.
The results of the ablation experiments conducted on the three datasets.
This project code is based on BasicSR(https://github.com/XPixelGroup/BasicSR).