Research map / Image generation / Image restoration
You Do Not Need Additional Priors or Regularizers in Retinex-Based Low-Light Image Enhancement
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 WNT-Xent loss (weighted negatives) to contrastive learning (ablation) | RFR_CL · LOL | PSNR: 25.33 -> 26.00 | Helped |
| add progressive learning strategy on top of WNT-Xent (ablation) | RFR_CL · LOL | PSNR: 26.00 -> 26.33 | Helped |
| add progressive learning strategy to self-knowledge distillation (ablation) | RFR_SD · LOL | PSNR: 23.21 -> 26.41 | Helped |
| replace element-wise Retinex product with learned synthesis module (end-to-end RFR vs. no Retinex decomposition) (ablation) | RFR · LOL | PSNR: 22.48 -> 26.33 | Helped |
| use synthesis module instead of element-wise product for prior Retinex methods (ablation) | RetinexNet, KinD++, RRDNet, URetinexNet · LOL | PSNR: RetinexNet 16.77->20.30; KinD++ 21.30->22.62; RRDNet 11.33->16.80; URetinexNet 21.33->23.18 | Helped |
| KinD with synthesis module vs original KinD (ablation) | KinD · LOL | PSNR: 20.87 -> 21.02 | No clear effect |
| contrastive learning / self-knowledge distillation regularizer-free Retinex training vs prior SOTA | RFR · LOL | PSNR: 26.41 vs LLFlow 25.19 | Helped |
| cross-dataset training on LOL, testing on VE-LOL with RFR | RFR · VE-LOL | PSNR: 26.22 vs LLFlow 23.85 | Helped |
Closest papers by meaning
- IniRetinex: Rethinking Retinex-type Low-Light Image Enhancer via Initialization Perspective (AAAI 2025 · similarity 0.92)
- MambaLLIE: Implicit Retinex-Aware Low Light Enhancement with Global-then-Local State Space (NeurIPS 2024 · similarity 0.88)
- Reti-Diff: Illumination Degradation Image Restoration with Retinex-based Latent Diffusion Model (ICLR 2025 · similarity 0.88)
- Zero-Shot Low-Light Image Enhancement via Latent Diffusion Models (AAAI 2025 · similarity 0.88)
- Interpretable Unsupervised Joint Denoising and Enhancement for Real-World low-light Scenarios (ICLR 2025 · similarity 0.87)
- Dehaze-RetinexGAN: Real-World Image Dehazing via Retinex-based Generative Adversarial Network (AAAI 2025 · similarity 0.86)
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.