Research map / Image generation / Image restoration
Low-Light Image Enhancement via Structure Modeling and Guidance
Xiaogang Xu, Ruixing Wang, Jiangbo Lu
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 appearance modeling module A (ablation) | SMG-LLIE · LOL-real | PSNR: 20.17 -> 24.62 | Improved |
| Add structure modeling module S (ablation) | SMG-LLIE · LOL-real | SSIM: 0.773 -> 0.867 | Improved |
| Use SAFE instead of a traditional StyleGAN encoder (ablation) | Structure model S · LOL-real | SSIM: 0.812 -> 0.867 | Improved |
| Add structure-guided feature synthesis (ablation) | SGEM · LOL-real | PSNR: 19.39 -> 24.62 | Improved |
| Use the proposed structure model instead of an encoder-decoder edge predictor (ablation) | SMG-LLIE · LOL-real | PSNR: 20.73 -> 24.62 | Improved |
| Add GAN loss to structure modeling (ablation) | Structure model S · LOL-real | SSIM: 0.812 -> 0.867 | Improved |
| Add the refinement and structure-guided enhancement to appearance module A (ablation) | SMG-LLIE · LOL-real | PSNR: 18.99 -> 24.62 | Improved |
| Add Gaussian noise to the input (variance ranges from 30 to 50) | SMG-LLIE · LOL-real | PSNR: 24.62 -> 24.15 | Hurt |
| Fine-tune structure modeling with extra AVA data | SMG-LLIE · LOL-real | PSNR: 24.62 -> 25.57 | Improved |
| Use segmentation maps instead of edge maps for structural representation | SMG-LLIE · LOL-real | PSNR: 24.62 -> 24.18 | Hurt |
| Use depth maps instead of edge maps for structural representation | SMG-LLIE · LOL-real | PSNR: 24.62 -> 23.89 | Hurt |
Closest papers by meaning
- Low-Light Image Enhancement via Generative Perceptual Priors (AAAI 2025 · similarity 0.88)
- Zero-Reference Low-Light Enhancement via Physical Quadruple Priors (CVPR 2024 · similarity 0.83)
- ZERO-IG: Zero-Shot Illumination-Guided Joint Denoising and Adaptive Enhancement for Low-Light Images (CVPR 2024 · similarity 0.83)
- URWKV: Unified RWKV Model with Multi-state Perspective for Low-light Image Restoration (CVPR 2025 · similarity 0.83)
- AGLLDiff: Guiding Diffusion Models Towards Unsupervised Training-free Real-world Low-light Image Enhancement (AAAI 2025 · similarity 0.83)
- Zero-Shot Low-Light Image Enhancement via Latent Diffusion Models (AAAI 2025 · 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.
