Research map / Image generation / Image restoration
UDC-SIT: A Real-World Dataset for Under-Display Cameras
Kyusu Ahn, Byeonghyun Ko, HyunGyu Lee, Chanwoo Park, Jaejin Lee
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 DFT phase loss to MSE in alignment (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 78.77 -> 98.22 | Improved |
| use DFT amplitude loss alone without MSE (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 78.77 -> 34.05 | Hurt |
| use DFT phase loss alone without MSE (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 78.77 -> 62.27 | Hurt |
| add amplitude loss to MSE without phase (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 78.77 -> 35.55 | Hurt |
| add amplitude loss to MSE plus phase (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 98.22 -> 86.13 | Mixed |
| apply image rotation in alignment (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 78.77 -> 43.83 | Hurt |
| use small weights 0.1 for frequency losses (ablation) | UDC-SIT alignment · UDC-SIT PCK alpha=0.01 | PCK: 78.77 -> 72.40 | Hurt |
| train restoration on real UDC-SIT instead of synthetic Feng-S | ECFNet · restoration PSNR | PSNR: 52.17 -> 28.22 | Hurt |
| train UDC-UNet on UDC-SIT instead of Feng-S | UDC-UNet · restoration SSIM | SSIM: 0.9933 -> 0.8637 | Hurt |
| use UDC-SIT instead of Feng-R pseudo-real alignment | dataset · PCK alpha=0.01 | PCK: 58.75 -> 97.26 | Improved |
| outdoor vs indoor scenes | ECFNet · UDC-SIT restoration | PSNR: 28.64 -> 26.66 | Hurt |
| night vs day scenes | ECFNet · UDC-SIT restoration | PSNR: 28.17 -> 27.06 | Hurt |
Closest papers by meaning
- UDC-VIT: A Real-World Video Dataset for Under-Display Cameras (ICCV 2025 · similarity 0.92)
- Panel-Specific Degradation Representation for Raw Under-Display Camera Image Restoration (ECCV 2024 · similarity 0.91)
- Decoupling Degradations with Recurrent Network for Video Restoration in Under-Display Camera (AAAI 2024 · similarity 0.91)
- Image Restoration for Under-Display Camera (CVPR 2021 · similarity 0.90)
- UltraLED: Learning to See Everything in Ultra-High Dynamic Range Scenes (NeurIPS 2025 · similarity 0.83)
- Correlation Matching Transformation Transformers for UHD Image Restoration (AAAI 2024 · similarity 0.83)
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.
