Dual-Scale Condition-Shift Editing for Robust Real-World Image Demoiréing

Key points

• Dedicated restoration networks trained on finite paired data struggle with this diversity, especially when low-frequency color banding and high-frequency stripes co-occur.1
• We present DemoiréEdit, a content-preserving editing framework that adapts the natural-image prior of a pretrained flow-matching editor to demoiréing.1
• 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.1
• 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.1

Sources (1)

• [1]Dual-Scale Condition-Shift Editing for Robust Real-World Image Demoiréing
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 19, 04:04 PM

  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.

Extractive summary: sentences quoted from the sources.