Research map / Video / Video segmentation and tracking
Dancing in the Dark: A Benchmark towards General Low-light Video Enhancement
Huiyuan Fu, Wenkai Zheng, Xicong Wang, Jiaxuan Wang, Heng Zhang, Huadong Ma
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 |
|---|---|---|---|
| adaptive light adjustment (hypothesis test on iteration count) instead of fixed T=10 (ablation) | LAN · DID test set | PSNR: 27.28 -> 29.01 | Improved |
| adaptive light adjustment instead of fixed T=10 (ablation) | LAN · SDSD test set | PSNR: 26.95 -> 27.25 | Improved |
| train on DID instead of SDSD | low-light enhancement models · user study on real phone videos | preference share: MBLLEN 96.3% prefer DID-trained | Improved |
| train on DID instead of SDSD for SDSDNet | SDSDNet · user study | preference share: SDSD 38.1% vs DID 42.9% | Mixed |
| LAN vs MBLLEN in user study | LAN · user study | preference share: Other 40.7% vs LAN 44.4% | Mixed |
| training on SDSD with imprecise alignment | enhancement models · real-captured low-light videos | blurriness of results | Hurt |
| train LAN on DID and evaluate on DID versus SDSD vs best baseline | LAN · DID and SDSD | PSNR: more than 3dB on DID, more than 1.9dB on SDSD | Improved |
Closest papers by meaning
- Seeing Dynamic Scene in the Dark: A High-Quality Video Dataset With Mechatronic Alignment (ICCV 2021 · similarity 0.89)
- RetinexMCNet: A Memory Controller Dominated Network for Low-Light Video Enhancement Based on Retinex (ICCV 2025 · similarity 0.88)
- Temporally Consistent Enhancement of Low-Light Videos via Spatial-Temporal Compatible Learning (IJCV 2024 · similarity 0.88)
- Unrolled Decomposed Unpaired Learning for Controllable Low-Light Video Enhancement (ECCV 2024 · similarity 0.88)
- NIR-assisted Video Enhancement via Unpaired 24-hour Data (ICCV 2023 · similarity 0.87)
- Towards Robust Event-guided Low-Light Image Enhancement: A Large-Scale Real-World Event-Image Dataset and Novel Approach (CVPR 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.
