Research map / Image generation / Super-resolution
Towards Efficient and Scale-Robust Ultra-High-Definition Image Demoiréing
Xin Yu, Peng Dai, Wenbo Li, Lan Ma, Jiajun Shen, Jia Li, Xiaojuan Qi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add pyramid context extraction with shared weights to baseline (ablation) | ESDNet SAM · UHDM | PSNR: 20.646 -> 21.176 | Improved |
| add pyramid context extraction with adaptive weights to baseline (ablation) | ESDNet SAM · UHDM | PSNR: 20.646 -> 21.300 | Improved |
| add cross-scale dynamic fusion to shared-weight pyramid (ablation) | ESDNet SAM · UHDM | PSNR: 21.176 -> 21.958 | Improved |
| add cross-scale dynamic fusion to adaptive pyramid (ablation) | ESDNet SAM · UHDM | PSNR: 21.300 -> 22.119 | Improved |
| compare stronger same-capacity baseline A+ with full model (ablation) | ESDNet · UHDM | PSNR: 20.860 -> 22.119 | Improved |
| replace L1 loss with proposed perceptual loss (ablation) | ESDNet · UHDM | PSNR: 21.489 -> 22.119 | Improved |
| replace L1 loss with proposed perceptual loss (ablation) | ESDNet · UHDM | LPIPS: 0.3330 -> 0.2551 | Improved |
| train DMCNN with proposed loss instead of L1 (ablation) | DMCNN · UHDM | LPIPS: 0.3764 -> 0.3089 | Improved |
| train MDDM with proposed loss instead of L1 (ablation) | MDDM · UHDM | LPIPS: 0.3409 -> 0.2986 | Improved |
| add SAM to DMCNN (ablation) | DMCNN · UHDM | PSNR: 19.914 -> 20.769 | Improved |
| add SAM to MDDM (ablation) | MDDM · UHDM | PSNR: 20.088 -> 20.883 | Improved |
| add SAM to MBCNN (ablation) | MBCNN · UHDM | PSNR: 21.414 -> 21.532 | Improved |
Closest papers by meaning
- UniDemoiré: Towards Universal Image Demoiréing with Data Generation and Synthesis (AAAI 2025 · similarity 0.90)
- Coarse-to-fine Disentangling Demoiréing Framework for Recaptured Screen Images (TPAMI 2023 · similarity 0.89)
- Self-Adaptively Learning to Demoiré from Focused and Defocused Image Pairs (NeurIPS 2020 · similarity 0.89)
- Learning Image Demoiréing from Unpaired Real Data (AAAI 2024 · similarity 0.87)
- From Zero to Detail: Deconstructing Ultra-High-Definition Image Restoration from Progressive Spectral Perspective (CVPR 2025 · similarity 0.85)
- UltraPixel: Advancing Ultra High-Resolution Image Synthesis to New Peaks (NeurIPS 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
