Research map / Image generation / Autoregressive and latent generation
GLARE: Low Light Image Enhancement via Generative Latent Feature based Codebook Retrieval
Han Zhou, Wei Dong, Xiaohong Liu, Shuaicheng Liu, Xiongkuo Min, Guangtao Zhai, Jun Chen
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 the adaptive feature transformation (AFT) module to SimGLARE (ablation) | GLARE · LOL-v1 | PSNR: 26.51 -> 27.35 | Improved |
| add dual-decoder architecture on top of AMB (ablation) | GLARE · LOL-v1 | PSNR: 26.60 -> 27.35 | Improved |
| use adaptive mix-up block instead of skip connection (ablation) | GLARE · LOL-v1 | PSNR: 26.88 -> 27.35 | Improved |
| train with NLL loss instead of L1 loss (ablation) | SimGLARE · LOL-v1 | PSNR: 25.69 -> 26.51 | Improved |
| use I-LNF instead of Transformer code-index prediction (ablation) | SimGLARE (L1) · LOL-v1 | PSNR: 25.12 -> 25.69 | Improved |
| add the I-LNF module (ablation) | SimGLARE (L1) · LOL-v1 | PSNR: 24.53 -> 25.69 (1.16 dB) | Improved |
| add the normal-light codebook prior (ablation) | SimGLARE (L1) · LOL-v1 | PSNR: 23.42 -> 25.69 | Improved |
| add perceptual loss to L1 in stage III (ablation) | GLARE · LOL-v1 | PSNR: 27.02 -> 27.09 | Improved |
| add SSIM loss to L1 + perceptual in stage III (ablation) | GLARE · LOL-v1 | PSNR: 27.09 -> 27.35 | Improved |
| use GLARE as low-light enhancement preprocessing for detection | YOLO-v3 on ExDark · ExDark detection | mAP: 76.32 -> 77.50 | Improved |
| GLARE versus competing methods on unpaired datasets | GLARE · LIME and NPE (NIQE) | NIQE: LIME 4.52 vs RFR 3.81; NPE 4.19 vs LL-SKF 4.08 | Mixed |
Later papers that compared with this method
2 later accepted papers report GLARE as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Bayesian Neural Networks for One-to-Many Mapping in Image Enhancement (AAAI 2026 · better on 4 of 4 shared results)
- Multinex: Lightweight Low-light Image Enhancement via Multi-prior Retinex (CVPR 2026 · better on 0 of 1 shared result)
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
- Learnable Feature Patches and Vectors for Boosting Low-light Image Enhancement without External Knowledge (ICCV 2025 · similarity 0.89)
- Low-Light Image Enhancement with Multi-Stage Residue Quantization and Brightness-Aware Attention (ICCV 2023 · similarity 0.89)
- Zero-Shot Low-Light Image Enhancement via Latent Diffusion Models (AAAI 2025 · similarity 0.88)
- Luminance-Aware Statistical Quantization: Unsupervised Hierarchical Learning for Illumination Enhancement (NeurIPS 2025 · similarity 0.87)
- Zero-Reference Low-Light Enhancement via Physical Quadruple Priors (CVPR 2024 · similarity 0.86)
- VQCNIR: Clearer Night Image Restoration with Vector-Quantized Codebook (AAAI 2024 · 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.
