Research map / Image generation / Autoregressive and latent generation
Unsupervised Exposure Correction
Ruodai Cui, Li Niu, Guosheng Hu
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 unsupervised radiometry-only training instead of supervised expert ground truth | UEC vs ECM · MSEC test | PSNR: 20.874 (ECM) vs 18.756 (UEC) | Hurt |
| train unsupervised UEC instead of supervised ECM, test on unseen LOL | exposure correction · LOL (trained on MSEC) | PSNR: 15.439 -> 18.571 | Improved |
| train ECM on Radiometry Correction Dataset instead of MSEC (ablation) | ECM · LOL | PSNR: 15.439 -> 17.537 | Improved |
| use UEC pixel-wise transformation instead of ECM on radiometry dataset | exposure correction · Radiometry Correction Dataset | SSIM (avg): 0.744 -> 0.868 | Improved |
| use UEC output instead of ECM output before edge detection | LDC edge detector · Radiometry Correction Dataset, edge detection | edge PSNR (avg): 16.312 -> 22.665 | Improved |
| choose a brighter reference image on MSEC (ablation) | UEC · MSEC | PSNR: 27.733 -> 28.097 across references | No clear effect |
| choose a brighter reference image on the Radiometry dataset (ablation) | UEC · Radiometry Correction Dataset | PSNR: 22.577 -> 15.138 | Hurt |
| use UEC small network instead of ECM | inference efficiency · 256x256 inference | CPU time: 149.39ms -> 6.38ms | Improved |
Closest papers by meaning
- Region-Aware Exposure Consistency Network for Mixed Exposure Correction (AAAI 2024 · similarity 0.85)
- Exposure-slot: Exposure-centric Representations Learning with Slot-in-Slot Attention for Region-aware Exposure Correction (CVPR 2025 · similarity 0.85)
- ECMamba: Consolidating Selective State Space Model with Retinex Guidance for Efficient Multiple Exposure Correction (NeurIPS 2024 · similarity 0.83)
- Real-Time Exposure Correction via Collaborative Transformations and Adaptive Sampling (CVPR 2024 · similarity 0.82)
- CostFilter-AD: Enhancing Anomaly Detection through Matching Cost Filtering (ICML 2025 · similarity 0.82)
- Rethinking No-reference Image Exposure Assessment from Holism to Pixel: Models, Datasets and Benchmarks (NeurIPS 2024 · similarity 0.82)
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.
