Research map / Image generation / Image restoration
Task-Decoupled Bezier Surface Constraint for Uneven Low-Light Image Enhancement
Xingxiang Zhou, Xiangdong Su, Haoran Zhang, Wei Chen, Guanglai Gao
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 decoupled branch structure (one encoder, two decoders) (ablation) | BSCNet baseline · LOLv1 | PSNR: 27.75 -> 28.05 | Improved |
| add Bezier surface constraint without branch structure (ablation) | BSCNet baseline · LOLv1 | PSNR: 27.75 -> 27.29 | Hurt |
| add Bezier surface constraint on top of branch structure (ablation) | BSCNet with branch · LOLv1 | PSNR: 28.05 -> 28.25 | Improved |
| add SFRM alone (ablation) | BSCNet baseline · LOLv1 | SSIM: 0.910 -> 0.922 | Improved |
| add SFRM alone (ablation) | BSCNet baseline · LOLv1 | PSNR: 27.75 -> 27.58 | No clear effect |
| add SFRM to branch + Bezier (ablation) | BSCNet · LOLv1 | PSNR: 28.25 -> 28.07 | Hurt |
| add SFRM to branch + Bezier (ablation) | BSCNet · LOLv1 | SSIM: 0.919 -> 0.923 | Improved |
| use BSCNet instead of Retinexformer | low-light enhancement · LOLv2-real | PSNR: 22.80 -> 28.33 | Improved |
| use BSCNet instead of PyDiff | low-light enhancement · FiveK | PSNR: 25.81 -> 27.95 | Improved |
| use BSCNet instead of Retinexformer | low-light enhancement, no-reference · DICM | NIQE: 3.86 -> 3.85 | No clear effect |
| use BSCNet instead of Retinexformer | low-light enhancement, no-reference · DICM | BRISQUE: 16.21 -> 21.17 | Hurt |
| use BSCNet instead of SNR-Net on brightness smoothness | low-light enhancement · LOLv1 | Bs: 0.084 -> 0.093 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Retinexformer — Retinexformer: One-stage Retinex-based Transformer for Low-light Image Enhancement (ICCV 2023) · better on 11 of 11 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Low-Light Image Enhancement with Multi-Stage Residue Quantization and Brightness-Aware Attention (ICCV 2023 · similarity 0.87)
- Learnable Feature Patches and Vectors for Boosting Low-light Image Enhancement without External Knowledge (ICCV 2025 · similarity 0.86)
- Ultra-High-Definition Low-Light Image Enhancement: A Benchmark and Transformer-Based Method (AAAI 2023 · similarity 0.86)
- Enhancing Low-Light Images: A Synthetic Data Perspective on Practical and Generalizable Solutions (AAAI 2025 · similarity 0.86)
- Zero-Shot Low-Light Image Enhancement via Latent Diffusion Models (AAAI 2025 · similarity 0.85)
- HVI: A New Color Space for Low-light Image Enhancement (CVPR 2025 · similarity 0.84)
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.
