Title:FPANet: Frequency-based Video Demoireing using Frame-level Post Alignment

Abstract:Interference between overlapping gird patterns creates moire patterns, degrading the visual quality of an image that captures a screen of a digital display device by an ordinary digital camera. Removing such moire patterns is challenging due to their complex patterns of diverse sizes and color distortions. Existing approaches mainly focus on filtering out in the spatial domain, failing to remove a large-scale moire pattern. In this paper, we propose a novel model called FPANet that learns filters in both frequency and spatial domains, improving the restoration quality by removing various sizes of moire patterns. To further enhance, our model takes multiple consecutive frames, learning to extract frame-invariant content features and outputting better quality temporally consistent images. We demonstrate the effectiveness of our proposed method with a publicly available large-scale dataset, observing that ours outperforms the state-of-the-art approaches, including ESDNet, VDmoire, MBCNN, WDNet, UNet, and DMCNN, in terms of the image and video quality metrics, such as PSNR, SSIM, LPIPS, FVD, and FSIM.