Research map / Video / Video generation
DreamStyle: A Unified Framework for Video Stylization
Mengtian Li, Jinshu Chen, Songtao Zhao, Wanquan Feng, Pengqi Tu, Qian He
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 token-specific LoRA (ablation) | DreamStyle · Style-image-guided stylization | CSD score: 0.413 -> 0.515 | Improved |
| add token-specific LoRA (ablation) | DreamStyle · Style-image-guided stylization | DINO score: 0.518 -> 0.526 | Improved |
| add SFT stage after CT training (ablation) | DreamStyle · Style-image-guided stylization | CSD score: 0.459 -> 0.515 | Improved |
| add CT stage before SFT training (ablation) | DreamStyle · Style-image-guided stylization | DINO score: 0.483 -> 0.526 | Improved |
| add CT stage before SFT training (ablation) | DreamStyle · Style-image-guided stylization | CSD score: 0.535 -> 0.515 | Hurt |
| add SFT stage after CT training (ablation) | DreamStyle · Style-image-guided stylization | DINO score: 0.547 -> 0.526 | Hurt |
| use DreamStyle instead of StyleMaster | style-image-guided stylization · Style-image-guided (T2V) | CSD score: 0.198 -> 0.532 | Improved |
| use DreamStyle instead of Runway for text-guided | text-guided stylization · Text-guided | DINO score: 0.504 -> 0.584 | Improved |
| use DreamStyle instead of Pixverse on image quality | text-guided stylization · Text-guided | Image quality: 0.746 -> 0.738 | Hurt |
| use DreamStyle instead of VACE for structure preservation | first-frame-guided stylization · First-frame-guided | DINO score: 0.716 -> 0.640 | Hurt |
| use DreamStyle instead of VideoX-Fun for style consistency | first-frame-guided stylization · First-frame-guided | CSD score: 0.766 -> 0.851 | Improved |
| drop video condition (T2V mode) from DreamStyle | DreamStyle · Style-image-guided | CSD score: 0.515 -> 0.532 | Improved |
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.
- StyleMaster — StyleMaster: Stylize Your Video with Artistic Generation and Translation (CVPR 2025) · better on 1 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
- FreeViS: Training-free Video Stylization with Inconsistent References (ICLR 2026 · similarity 0.91)
- EchoStyle: Unlocking High-Fidelity Video Stylization with Reverse Data Synthesis (ECCV 2026 · similarity 0.90)
- StyleMaster: Stylize Your Video with Artistic Generation and Translation (CVPR 2025 · similarity 0.89)
- Semantix: An Energy-guided Sampler for Semantic Style Transfer (ICLR 2025 · similarity 0.87)
- Generative Human Motion Stylization in Latent Space (ICLR 2024 · similarity 0.86)
- Local Consistency Guidance: Personalized Stylization Method of Face Video (Student Abstract) (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.
