Research map / Video / Video segmentation and tracking
There and Back Again: A Flexible-Frame Transformer for Multi-Exposure Fusion
Lishen Qu, Yao Liu, shihao zhou, Jie Liang, Hui Zeng, Lei Zhang, Jufeng Yang
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 RSSM in place of CNN embedding (ablation) | Restormer-style baseline · Kalantari (3 frames) | PSNR: 27.082 -> 27.565 | Improved |
| add EAHA in place of MDTA (ablation) | Restormer-style baseline · Kalantari (3 frames) | PSNR: 27.082 -> 27.457 | Improved |
| add AFFN in place of GDFN (ablation) | Restormer-style baseline · Kalantari (3 frames) | PSNR: 27.082 -> 27.426 | Improved |
| add RSSM to EAHA+AFFN model (ablation) | FreeMEF · Kalantari (3 frames) | PSNR: 27.982 -> 28.418 | Improved |
| add EAHA to RSSM+AFFN model (ablation) | FreeMEF · Kalantari (3 frames) | PSNR: 27.837 -> 28.418 | Improved |
| add AFFN to RSSM+EAHA model (ablation) | FreeMEF · Kalantari (3 frames) | PSNR: 28.073 -> 28.418 | Improved |
| recurrent fusion vs fixed-input parallel baselines on SICE 2-frame | FreeMEF · SICE 2-frame (trained on Kalantari) | PSNR: 15.339 -> 17.087 | Improved |
| recurrent fusion vs fixed-input baselines on SICE 5-frame | FreeMEF · SICE 5-frame (trained on Kalantari) | PSNR: 21.153 -> 22.269 | Improved |
| fuse 5 frames brightest to darkest (II) instead of other orders (ablation) | FreeMEF · SICE 5-frame | PSNR | Hurt |
| fuse 5 frames in orders I, III or IV (ablation) | FreeMEF · SICE 5-frame | PSNR | No clear effect |
Closest papers by meaning
- Hybrid-Supervised Dual-Search: Leveraging Automatic Learning for Loss-Free Multi-Exposure Image Fusion (AAAI 2024 · similarity 0.88)
- A Unified Solution to Video Fusion: From Multi-Frame Learning to Benchmarking (NeurIPS 2025 · similarity 0.87)
- Ultra-High-Definition Dynamic Multi-Exposure Image Fusion via Infinite Pixel Learning (AAAI 2025 · similarity 0.87)
- Every SAM Drop Counts: Embracing Semantic Priors for Multi-Modality Image Fusion and Beyond (CVPR 2025 · similarity 0.87)
- Motion-adaptive Transformer for Event-based Image Deblurring (AAAI 2025 · similarity 0.86)
- Towards Robust Event-guided Low-Light Image Enhancement: A Large-Scale Real-World Event-Image Dataset and Novel Approach (CVPR 2024 · 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.
