Research map / Image generation / Image restoration
HDR Imaging With Spatially Varying Signal-to-Noise Ratios
Yiheng Chi, Xingguang Zhang, Stanley H. Chan
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 |
|---|---|---|---|
| Use SV-HDR instead of competing HDR fusion methods | Synthetic HDR fusion · Synthetic HDR, 0.4–0.05 lux | PSNR (dB): SV-HDR: 41.75, 41.61, 41.36, 40.98; strongest baseline AHDRNet: 36.46, 36.61, 36.77, 36.87 | Improved |
| Use SW-MSA in both encoder and decoder rather than the tested alternative attention pairings (ablation) | SV-HDR · Kalantari test dataset, τ=4 | PSNR (dB): 40.28; alternatives: 25.48, 25.48, 39.68, 38.91, 39.33, 39.37, 39.39, 39.90 | Improved |
| Use MLP rather than LeFF, DeLeFF, or GDFN (ablation) | SV-HDR · Kalantari test dataset, τ=4 | PSNR (dB): 40.34 vs 39.65, 39.88, 40.09 | Improved |
| Use deformable convolutions instead of standard convolutions in Expo-Share blocks (ablation) | SV-HDR · Synthetic HDR across illuminance levels | PSNR | Improved |
| Increase synthetic input photon level from 10 to 100 photons per pixel (ablation) | Existing HDR fusion algorithms · Synthetic HDR fusion | Reconstruction quality | Improved |
| Use a denoiser bank with finer luminance subranges instead of one denoiser per exposure (ablation) | Toy denoising followed by HDR fusion · Synthetic HDR fusion | Reconstruction quality | Improved |
Closest papers by meaning
- Single Shot HDR Recovery via a Video Diffusion Prior (NeurIPS 2026 · similarity 0.87)
- UltraFusion: Ultra High Dynamic Imaging using Exposure Fusion (CVPR 2025 · similarity 0.86)
- Color Event Enhanced Single-Exposure HDR Imaging (AAAI 2024 · similarity 0.86)
- EvHDR-NeRF: Building High Dynamic Range Radiance Fields with Single Exposure Images and Events (AAAI 2025 · similarity 0.86)
- High Dynamic Range Novel View Synthesis with Single Exposure (ICML 2025 · similarity 0.86)
- HDR-GS: Efficient High Dynamic Range Novel View Synthesis at 1000x Speed via Gaussian Splatting (NeurIPS 2024 · similarity 0.86)
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.
