Research map / Image generation / Image restoration
Graphics2RAW: Mapping Computer Graphics Images to Sensor RAW Images
Donghwan Seo, Abhijith Punnappurath, Luxi Zhao, Abdelrahman Abdelhamed, Sai Kiran Tedla, Sanguk Park, Jihwan Choe, Michael S. Brown
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 |
|---|---|---|---|
| map graphics to RAW via illumination-conditioned CST interpolation and Gaussian illuminant sampling instead of UPI sampling | Graphics2RAW · NUS illuminant estimation | mean angular error: 6.26 -> 4.21 | Improved |
| use Graphics2RAW instead of UPI for synthetic training data | Neural ISP · Nighttime dataset | PSNR: 36.43 -> 38.10 | Improved |
| use Graphics2RAW instead of UPI for synthetic training data | Restormer denoiser · Nighttime RAW denoising ISO 3200 | PSNR: 47.51 -> 48.16 | Improved |
| use Graphics2RAW instead of UPI for synthetic training data | Restormer denoiser · Nighttime RAW denoising ISO 1600 | PSNR: 49.05 -> 49.37 | Improved |
| use Graphics2RAW instead of UPI for synthetic training data | Restormer denoiser · Nighttime RAW denoising ISO 1600 | SSIM: 0.9904 -> 0.9911 | No clear effect |
| use Graphics2RAW instead of UPI for synthetic training data | Neural ISP · Nighttime dataset | Delta E: 2.907 -> 2.301 | Improved |
| train on synthetic Graphics2RAW data rather than real data | Neural ISP · Nighttime dataset | PSNR: 38.10 vs 38.32 (real) | No clear effect |
Closest papers by meaning
- Learning sRGB-to-Raw-RGB De-Rendering With Content-Aware Metadata (CVPR 2022 · similarity 0.86)
- Model-Based Image Signal Processors via Learnable Dictionaries (AAAI 2022 · similarity 0.86)
- R4-CGQA: Retrieval-based Vision Language Models for Computer Graphics Image Quality Assessment (CVPR 2026 · similarity 0.84)
- Lighting Every Darkness with 3DGS: Fast Training and Real-Time Rendering for HDR View Synthesis (NeurIPS 2024 · similarity 0.84)
- RAWild: Toward Sensor-Agnostic RAW Object Detection via Physics-Guided Curve and Grid Modeling (NeurIPS 2026 · similarity 0.83)
- ReRAW: RGB-to-RAW Image Reconstruction via Stratified Sampling for Efficient Object Detection on the Edge (CVPR 2025 · 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.
