Research map / Architectures and efficiency / CNN design
Improving Dynamic HDR Imaging with Fusion Transformer
Rufeng Chen, Bolun Zheng, Hua Zhang, Quan Chen, Chenggang Yan, Gregory Slabaugh, Shanxin Yuan
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 HFT instead of the strongest reported baseline | HDR reconstruction · Kalantari’s dataset | PSNR-µ: 44.28 -> 44.45 | Improved |
| Use HFT instead of HDRRNN | HDR reconstruction · Prabhakar’s dataset | PSNR-µ: 39.03 -> 40.12 | Improved |
| Add SFA feature alignment (ablation) | HFT · Kalantari’s dataset | PSNR-µ: 44.04 -> 44.45 | Improved |
| Add HDR Transformer at the smallest scale (ablation) | HFT · Kalantari’s dataset | PSNR-µ: 44.26 -> 44.45 | Improved |
| Add CADB blocks to reconstruction (ablation) | HFT · Kalantari’s dataset | PSNR-µ: 44.26 -> 44.45 | Improved |
| Add SFA feature alignment (ablation) | HFT · Prabhakar’s dataset | PSNR-µ: 39.53 -> 40.15 | Improved |
| Add HDR Transformer at the smallest scale (ablation) | HFT · Prabhakar’s dataset | PSNR-µ: 39.81 -> 40.15 | Improved |
| Add CADB blocks to reconstruction (ablation) | HFT · Prabhakar’s dataset | PSNR-µ: 39.53 -> 40.15 | Improved |
| Use ASL instead of L1 loss with 3 CADBs (ablation) | HFT · Kalantari’s dataset | PSNR-µ: 43.58 -> 44.45 | Improved |
| Use 3 CADBs instead of 2 with ASL (ablation) | HFT · Kalantari’s dataset | PSNR-µ: 43.18 -> 44.45 | Improved |
| Use 3 CADBs instead of 1 with ASL (ablation) | HFT · Kalantari’s dataset | PSNR-µ: 44.35 -> 44.45 | Improved |
Closest papers by meaning
- Selective TransHDR: Transformer-Based Selective HDR Imaging Using Ghost Region Mask (ECCV 2022 · similarity 0.92)
- Single-Image HDR Reconstruction by Learning to Reverse the Camera Pipeline (CVPR 2020 · similarity 0.91)
- Single Shot HDR Recovery via a Video Diffusion Prior (NeurIPS 2026 · similarity 0.89)
- Self-Supervised High Dynamic Range Imaging with Multi-Exposure Images in Dynamic Scenes (ICLR 2024 · similarity 0.88)
- Generating Content for HDR Deghosting from Frequency View (CVPR 2024 · similarity 0.88)
- High Dynamic Range Novel View Synthesis with Single Exposure (ICML 2025 · similarity 0.88)
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.
