Research map / Image generation / Image restoration
Selective TransHDR: Transformer-Based Selective HDR Imaging Using Ghost Region Mask
Jou Won Song, Ye-In Park, Kyeongbo Kong, Jaeho Kwak, Suk-Ju Kang
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 CNN for non-ghost region and transformer for ghost region instead of transformer only (ablation) | Selective TransHDR · Kalantari test | PSNR-mu: 43.3244 -> 44.0981 | Improved |
| use transformer for ghost region instead of CNN only (ablation) | Selective TransHDR · Kalantari test | PSNR-mu: 43.7274 -> 44.0981 | Improved |
| add ghost region mask guidance (ablation) | Selective TransHDR · Kalantari test | PSNR-mu: 43.8874 -> 44.0981 | Improved |
| add ghost region mask guidance (ablation) | Selective TransHDR · Kalantari test | HDR-VDP-2: 62.9245 -> 63.3721 | Improved |
| add ghost region mask guidance (ablation) | Selective TransHDR · Kalantari test | PSNR-L: 41.3422 -> 41.7021 | Improved |
| selective transformer vs CNN-only on HDR-VDP-2 (ablation) | Selective TransHDR · Kalantari test | HDR-VDP-2: 62.7231 -> 63.3721 | Improved |
| selective transformer vs transformer-only on HDR-VDP-2 (ablation) | Selective TransHDR · Kalantari test | HDR-VDP-2: 60.6124 -> 63.3721 | Improved |
| CNN-only beats transformer-only in non-ghost region (ablation) | base model · Kalantari test | PSNR-mu: 43.3244 -> 43.7274 | Improved |
| proposed vs best prior (HDRGAN) in ghost regions | Selective TransHDR · Kalantari test (ghost regions) | PSNR-mu: 41.2814 -> 41.6714 | Improved |
| proposed vs best prior (HDRGAN) on full image | Selective TransHDR · Kalantari test | PSNR-mu: 43.9220 -> 44.0981 | Improved |
| proposed vs HDRGAN on PSNR-L | Selective TransHDR · Kalantari test | PSNR-L: 41.5720 -> 41.7021 | Improved |
Closest papers by meaning
- Improving Dynamic HDR Imaging with Fusion Transformer (AAAI 2023 · similarity 0.92)
- Ghost-Free High Dynamic Range Imaging with Context-Aware Transformer (ECCV 2022 · similarity 0.90)
- Generating Content for HDR Deghosting from Frequency View (CVPR 2024 · similarity 0.87)
- Self-Supervised High Dynamic Range Imaging with Multi-Exposure Images in Dynamic Scenes (ICLR 2024 · similarity 0.86)
- Single Shot HDR Recovery via a Video Diffusion Prior (NeurIPS 2026 · similarity 0.86)
- GaussHDR: High Dynamic Range Gaussian Splatting via Learning Unified 3D and 2D Local Tone Mapping (CVPR 2025 · similarity 0.85)
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.
