Research map / Image generation / Image restoration
Low-Light Image Enhancement with Normalizing Flow
Yufei Wang, Renjie Wan, Wenhan Yang, Haoliang Li, Lap-Pui Chau, Alex Kot
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 conditional normalizing flow with NLL training instead of L1 training (ablation) | LLFlow · LOL | PSNR: 22.68 -> 25.19 | Improved |
| Use conditional normalizing flow with NLL training instead of L1 training (ablation) | LLFlow · LOL | SSIM: 0.8391 -> 0.9252 | Improved |
| Use conditional normalizing flow with NLL training instead of L1 training (ablation) | LLFlow · LOL | LPIPS: 0.2038 -> 0.1131 | Improved |
| Add the conditional invertible flow to the encoder-only model (ablation) | LLFlow · LOL | PSNR: 21.90 -> 25.19 | Improved |
| Add the illumination-invariant color map prior (ablation) | LLFlow · LOL | PSNR: 24.46 -> 25.19 | Improved |
| Add the illumination-invariant color map prior (ablation) | LLFlow · LOL | SSIM: 0.9235 -> 0.9252 | Improved |
| Add the illumination-invariant color map prior (ablation) | LLFlow · LOL | LPIPS: 0.1146 -> 0.1131 | Improved |
| Set color-map selection probability p to 0.5 rather than 0.2 (ablation) | LLFlow · LOL | PSNR: 25.19 -> 24.85 | Hurt |
| Set color-map selection probability p to 0.5 rather than 0.2 (ablation) | LLFlow · LOL | SSIM: 0.9252 -> 0.9232 | Hurt |
| Set color-map selection probability p to 0.5 rather than 0.2 (ablation) | LLFlow · LOL | LPIPS: 0.1131 -> 0.1192 | Hurt |
| Use latent z values with different offsets (ablation) | LLFlow · LOL | Brightness | Improved |
| Use the learned NLL rather than pixel-wise L1 as the image-quality likelihood (ablation) | LLFlow · LOL test image perturbations | NLL: -5.95; -6.15; 4.84 | Improved |
Closest papers by meaning
- Empowering Low-Light Image Enhancer through Customized Learnable Priors (ICCV 2023 · similarity 0.89)
- Zero-Shot Low-Light Image Enhancement via Latent Diffusion Models (AAAI 2025 · similarity 0.88)
- Zero-Reference Low-Light Enhancement via Physical Quadruple Priors (CVPR 2024 · similarity 0.87)
- Luminance-Aware Statistical Quantization: Unsupervised Hierarchical Learning for Illumination Enhancement (NeurIPS 2025 · similarity 0.87)
- DPLUT: Unsupervised Low-light Image Enhancement with Lookup Tables and Diffusion Priors (AAAI 2025 · similarity 0.87)
- Interpretable Unsupervised Joint Denoising and Enhancement for Real-World low-light Scenarios (ICLR 2025 · similarity 0.87)
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.
