Research map / Video / Video generation
Tuning-free Visual Effect Transfer across Videos
Maxwell Jones, Rameen Abdal, Or Patashnik, Ruslan Salakhutdinov, Sergey Tulyakov, Jun-Yan Zhu, Kuan-Chieh Wang
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 reference effect video conditioning (ablation) | RefVFX · Neural V2V user study | Overall Match win rate (%): 67.2 ± 2.9 | Improved |
| add reference effect video conditioning (ablation) | RefVFX · Neural V2V user study | Reference Video Adherence win rate (%): 57.5 ± 2.8 | Improved |
| add reference effect video conditioning (ablation) | RefVFX · Code-based V2V user study | Overall Match win rate (%): 61.3 ± 3.9 | Improved |
| add reference effect video conditioning (ablation) | RefVFX · I2V user study | Overall Match win rate (%): 61.8 ± 3.4 | Improved |
| add reference effect video conditioning (ablation) | RefVFX · Neural V2V user study | Input Video Adherence win rate (%): 51.3 ± 3.1 | No clear effect |
| use RefVFX instead of Lucy Edit | RefVFX · Neural V2V user study | Reference Video Adherence win rate (%): 62.3 ± 2.9 | Improved |
| use RefVFX instead of VACE Pose | RefVFX · Neural V2V user study | Input Video Adherence win rate (%): 53.7 ± 2.2 | Improved |
| use RefVFX instead of VACE Depth | RefVFX · Code-based V2V user study | Reference Video Adherence win rate (%): 65.7 ± 5.4 | Improved |
| use RefVFX instead of Wan2.1 prompt-only | RefVFX · I2V user study | Reference Video Adherence win rate (%): 57.1 ± 2.4 | Improved |
| use RefVFX instead of Wan 2.1 | RefVFX · I2V similarity | Ref Sim (VideoPrism): 0.7230 -> 0.7378 | Improved |
| use RefVFX instead of Wan 2.1 | RefVFX · I2V similarity | First Frame Sim (CLIP): 0.7911 -> 0.7698 | Hurt |
| use RefVFX instead of Lucy Edit | RefVFX · Neural V2V similarity | Ref Sim (VideoPrism): 0.6852 -> 0.7014 | Improved |
Closest papers by meaning
- EffectMaker: Unifying Reasoning and Generation for Customized Visual Effect Creation (CVPR 2026 · similarity 0.87)
- VINCIE-NExT: Unlocking Video Editing from Images via In-Context Modeling (NeurIPS 2026 · similarity 0.86)
- Rebalancing Reference Frame Dominance to Improve Motion in Image-to-Video Models (NeurIPS 2026 · similarity 0.86)
- VideoDirector: Precise Video Editing via Text-to-Video Models (CVPR 2025 · similarity 0.85)
- FlowMo: Variance-Based Flow Guidance for Coherent Motion in Video Generation (NeurIPS 2025 · similarity 0.85)
- TransPixeler: Advancing Text-to-Video Generation with Transparency (CVPR 2025 · similarity 0.85)
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.
