用于鲁棒真实世界图像去摩尔纹的双尺度条件偏移编辑
Dual-Scale Condition-Shift Editing for Robust Real-World Image Demoiréing
AI总结：
本文提出DemoiréEdit框架，结合双尺度条件偏移策略，构建DailyMoiré130基准，在去摩尔纹任务上取得优于基线的性能，获71.3%质量偏好。
摘要：
屏幕重拍图像的鲁棒去摩尔纹仍具挑战性：摩尔纹在不同显示-相机对及拍摄条件下，频率、颜色和空间支撑均存在差异。在有限配对数据上训练的专用恢复网络难以应对这种多样性，尤其是当低频色带与高频条纹共存时。本文提出DemoiréEdit，一种保留内容的编辑框架，将预训练的流匹配编辑器的自然图像先验适配到去摩尔纹任务。首先，端点监督训练目标在速度匹配基础上补充像素、感知和频域恢复损失；其次，双尺度条件偏移策略在低分辨率下执行全局校正，再通过将去噪条件从第一阶段输出偏移至原始输入，在全分辨率下细化图像以恢复细节。本文还构建了DailyMoiré130基准，包含130张不同显示设备下的日常手机拍摄摩尔纹图像，涵盖低频色带、精细高频条纹及混合频率干扰。在该基准上，本文方法实现最低残留摩尔纹面积，在31名参与者参与的盲选7最佳研究中获得71.3%的整体质量偏好；在配对FHDMi/UHDM数据集上，相比各指标最佳基线，残留摩尔纹置信度分别降低47.1%/9.3%，MUSIQ分别提升4.3%/2.0%。
英文摘要：
Robust demoiréing of screen-recaptured images remains challenging: moiré varies in frequency, color, and spatial support across display-camera pairs and capture conditions. Dedicated restoration networks trained on finite paired data struggle with this diversity, especially when low-frequency color banding and high-frequency stripes co-occur. We present DemoiréEdit, a content-preserving editing framework that adapts the natural-image prior of a pretrained flow-matching editor to demoiréing. First, an endpoint-supervised training objective supplements velocity matching with pixel, perceptual, and frequency-domain restoration losses. Second, a dual-scale condition-shift strategy performs global correction at low resolution, then refines the image at full resolution by shifting the denoising condition from the first-stage output to the original input for detail recovery. We also introduce DailyMoiré130, a benchmark of challenging moiré in everyday screen photography, comprising 130 smartphone captures across diverse displays with low-frequency color banding, fine high-frequency stripes, and mixed-frequency interference. On the challenging DailyMoiré130 benchmark, our method achieves the lowest residual moiré area and receives 71.3% overall-quality preference in a blind Best-of-7 study with 31 participants. We also evaluate on paired FHDMi/UHDM, reducing residual moiré confidence by 47.1%/9.3% and increasing MUSIQ by 4.3%/2.0% over the best baseline for each metric.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
