Video Demoireing

Video demoireing aims to remove moiré patterns—unwanted visual artifacts caused by interference between repetitive patterns—from videos, improving image quality. Recent research focuses on developing deep learning models, often employing frequency-domain processing alongside spatial and temporal analysis, to effectively remove these patterns while maintaining temporal consistency. These models often leverage techniques like state-space models, directional transforms, and bilateral learning to achieve superior performance compared to previous methods, as measured by metrics like PSNR and visual quality assessments. This work is significant for enhancing the quality of videos captured from screens, impacting applications ranging from video conferencing to professional filmmaking.