Research map / Video / Video generation
LightCtrl: Training-free Controllable Video Relighting
Yizuo Peng, Xuelin Chen, Kai Zhang, Xiaodong Cun
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 |
|---|---|---|---|
| use dynamic cut-off (Cut-off-lbd) normal injection instead of direct 0.5+0.5 fusion in GAR (ablation) | LightCtrl GAR module · Light-trajectory video relighting | AQ: 0.5892 -> 0.6114 | Improved |
| use dynamic cut-off (Cut-off-lbd) normal injection instead of direct 0.5+0.5 fusion in GAR (ablation) | LightCtrl GAR module · Light-trajectory video relighting | FVD: 1010.9 -> 993.1 | Improved |
| use time-varying cut-off instead of fixed Cut-off-0.25 filter (ablation) | LightCtrl GAR module · Light-trajectory video relighting | AQ: 0.6097 -> 0.6114 | Improved |
| use time-varying cut-off instead of fixed Cut-off-0.25 filter (ablation) | LightCtrl GAR module · Light-trajectory video relighting | FVD: 998.7 -> 993.1 | Improved |
| inject decoded frames plus first-step residual instead of raw frames in LMI (ablation) | LightCtrl LMI module · Light-trajectory video relighting | AQ: 0.5946 -> 0.6114 | Improved |
| inject decoded frames plus first-step residual instead of raw frames in LMI (ablation) | LightCtrl LMI module · Light-trajectory video relighting | FVD: 1037.4 -> 993.1 | Improved |
| add progressive fusion (ablation) | LightCtrl · Temporal consistency | CLIP-Score: 0.9200 -> 0.9634 | Improved |
| add SDEdit noise 0.6 to IC-Light for smoothing | IC-Light · Light-trajectory video relighting | FVD: 1018.5 -> 1630.9 | Hurt |
| use normal maps instead of delight images as GAR feature (ablation) | LightCtrl GAR module · Qualitative comparison | visual quality | Improved |
| use LightCtrl instead of LAV-Traj under static light | video relighting · Static light relighting | PSNR_light: 16.630 -> 17.494 | Improved |
Closest papers by meaning
- UniLumos: Fast and Unified Image and Video Relighting with Physics-Plausible Feedback (NeurIPS 2025 · similarity 0.91)
- Lux Post Facto: Learning Portrait Performance Relighting with Conditional Video Diffusion and a Hybrid Dataset (CVPR 2025 · similarity 0.90)
- UniRelight: Learning Joint Decomposition and Synthesis for Video Relighting (NeurIPS 2025 · similarity 0.90)
- IllumiCraft: Unified Geometry and Illumination Diffusion for Controllable Video Generation (NeurIPS 2025 · similarity 0.90)
- Comprehensive Relighting: Generalizable and Consistent Monocular Human Relighting and Harmonization (CVPR 2025 · similarity 0.89)
- LightIt: Illumination Modeling and Control for Diffusion Models (CVPR 2024 · similarity 0.88)
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.
