Research map / Image generation / GANs
DeepExposure: Learning to Expose Photos with Asynchronously Reinforced Adversarial Learning
Runsheng Yu, Wenyu Liu, Yasen Zhang, Zhi Qu, Deli Zhao, Bo Zhang
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 |
|---|---|---|---|
| DeepExposure I vs Exposure (RL baseline) with local exposures | DeepExposure · MIT-Adobe FiveK test | MSE: 97.99 -> 95.44 | Improved |
| DeepExposure I vs Exposure | DeepExposure · MIT-Adobe FiveK test | PSNR: 28.27 -> 28.38 | Improved |
| learned RL exposure vs empirically crafted exposure fusion (FI) | DeepExposure · MIT-Adobe FiveK test | MSE: 105.2 -> 95.44 | Improved |
| unpaired adversarial training vs paired method DPED | DeepExposure · MIT-Adobe FiveK test | PSNR: 28.20 -> 28.38 | Improved |
| use discriminator as value function (DE II) instead of separate value network (DE I) (ablation) | DeepExposure · MIT-Adobe FiveK test | MSE: 95.44 -> 96.42 | Hurt |
| use discriminator as value function (DE II) instead of separate value network (DE I) (ablation) | DeepExposure · MIT-Adobe FiveK test | PSNR: 28.38 -> 28.33 | Hurt |
| use discriminator as value function (DE II) instead of separate value network (DE I) (ablation) | DeepExposure · MIT-Adobe FiveK test | training time to converge: 320 min -> 280 min | Improved |
Closest papers by meaning
- Learning Multi-Scale Photo Exposure Correction (CVPR 2021 · similarity 0.87)
- Learning to Control Camera Exposure via Reinforcement Learning (CVPR 2024 · similarity 0.86)
- Neural Photo Editing with Introspective Adversarial Networks (ICLR 2017 · similarity 0.85)
- Region-Aware Exposure Consistency Network for Mixed Exposure Correction (AAAI 2024 · similarity 0.83)
- Nearly Zero-Cost Protection Against Mimicry by Personalized Diffusion Models (CVPR 2025 · similarity 0.83)
- Learning Neural Exposure Fields for View Synthesis (NeurIPS 2025 · similarity 0.83)
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.
